← back to all models ← RevenueBench home
elorian

elorian/glimpse

Mean 0.990 · 38/40 perfect tests · $2.71 total · modalities in: text, image, video · out: text · each card: the prompt → the correct answer (gold) → this model's actual answer

What do the modalities mean?

Modalities in means what you can send this model: text, images, files, video, or audio. Modalities out means what it can send back. Bench tasks feed text extracts, so modalities do not change the score. They are here so you know what this model can handle in real work.

data-crm 5 tests
stage-hygiene-auditPASS 1.00
Plain-English question: How much of the open DS2+ pipeline is missing a substantive why-buys, and is anything missing DS1 timestamps or out of order?
Prompt — what the model saw
Audit stage hygiene on the open pipeline below. Rules: a deal at DS2 or later must have a substantive why-buys (15+ characters, not a placeholder such as tbd, n/a, see notes). A why_buys value of the form LONG(n) means the real text was present but truncated for display; treat it as present and substantive. Every deal should have a DS1 entry timestamp. A deal whose stage entry timestamps are out of order regressed (an earlier-stage timestamp is later than a later-stage timestamp). Report: percent of DS2+ deals non-compliant on why-buys overall and by owner (use the owners table for names), count of deals with no DS1 timestamp, count that regressed, and three process fixes. State definitions before numbers.
cost $0.2323 · 371s · in 82,287 / out 53,664 tok
✓ noncomp_pct✓ no_ds1_0✓ regressed_0✓ per_owner_breakdown✓ definitions_first✓ three_fixes

✓ Correct Answer (ground truth)

Plain-English answer: 81.5% of the 124 DS2+ deals are missing a substantive why-buys — Alex Franklin 48, Cole Ingram 19, Dana Mercer 18, Bryce Harmon 15, Farid Osman 1 — and every deal has a DS1 timestamp with none regressed.
The answer in one line: n_deals = 156; ds2plus = 124; noncomp_pct = 81.5; bryce_noncomp = 15; no_ds1 = 0; regressed = 0 (+1 more below)
n_deals156
bryce_noncomp15
ds2plus124
no_ds10
noncomp_pct81.5
regressed0
per_owner_noncompAlex Franklin=48 of 61 (78.7%); Cole Ingram=19 of 22 (86.4%); Dana Mercer=18 of 22 (81.8%); Bryce Harmon=15 of 18 (83.3%); Farid Osman=1 of 1 (100%)
graded checksnoncomp_pct · no_ds1_0 · regressed_0 · per_owner_breakdown · definitions_first · three_fixes

Model's actual answer

show the model's full answer
Definitions first.
- DS2+ = stage in {DS2, DS3, DS4, DS5}: 124 of 156 open deals (DS2 39 + DS3 61 + DS4 14 + DS5 10).
- Substantive why_buys = non-empty text of 15+ characters, not a placeholder (tbd, n/a, see notes). A value written as LONG(n) means real text truncated for display: counted as present and substantive.
- No DS1 timestamp = t_ds1 empty.
- Regressed = any earlier stage-entry timestamp later than a later stage-entry timestamp among the timestamps present (t_ds1..t_ds5 checked pairwise on non-empty fields).

Why-buys compliance (DS2+ only)
- Overall: 101 of 124 non-compliant = 81.5%. All 101 are empty fields; all 23 compliant values are LONG(n). Zero placeholder-style short values exist, so no placeholder exclusions were needed.
- By owner (non-compliant / DS2+ deals):
  - Alex Franklin: 48/61 = 78.7%
  - Cole Ingram: 19/22 = 86.4%
  - Dana Mercer: 18/22 = 81.8%
  - Bryce Harmon: 15/18 = 83.3%
  - Farid Osman: 1/1 = 100.0%
  - Elena Sinclair: 0 DS2+ deals (only DS1 deals) — not applicable.

DS1 timestamps
- Deals with no DS1 timestamp: 0 of 156. Every deal, including all 32 DS1 deals, has t_ds1 populated.

Stage-order integrity
- Deals that regressed: 0. All present timestamp sequences are non-decreasing. Gaps exist (e.g., DS5 deals missing t_ds4/t_ds5 such as Deal-403845, Deal-547B2B), but gaps are not regressions and are not counted here. (Note: a handful of deals have t_ds1 one day before create_date, e.g. Deal-927338, Deal-DCA846, Deal-D8ABF7, Deal-819506; that is a create-date/entry-date mismatch, not a stage regression, and falls outside the stated rules.)

Three process fixes
1. Required-field gate at DS2: make why_buys mandatory (15+ chars) before a deal can advance past DS2, with a blacklist on tbd/n/a/see notes and a one-time backfill sprint for the 101 open DS2+ deals with empty why_buys, listed by alias to each owner.
2. Automate stage-entry stamps: workflow stamps t_ds1 on deal creation and t_dsN on each stage advance, with monotonicity enforced (block a stage change if the current stage's timestamp is missing or earlier than the previous stage's). This keeps DS1 coverage and zero-regression status from degrading as volume grows.
3. Weekly pipeline hygiene monitor: automated check that reports (a) DS2+ deals with empty or placeholder why_buys and (b) any out-of-order or missing stage timestamps, sent to reps and managers with deal aliases; track owner-level non-compliance as a forecast-hygiene metric alongside pipeline coverage.

⚠️ File-mutation verifier: 1 file edit(s) FAILED this turn despite any wording above that may suggest otherwise. Run `git status` or `read_file` to confirm what actually landed.
  • `/Users/amaniphipps/.hermes/cache/scratch/deals_open.csv` — [write_file] Refusing to overwrite `/Users/amaniphipps/.hermes/cache/scratch/deals_open.csv`: `/Users/amaniphipps/.hermes/cache/scratch/deals_open.csv` exists but this task has not seen its full c…
owner-resolutionPASS 1.00
Plain-English question: Every open deal has a blank owner name — resolve each one to a real person using the owners table.
Prompt — what the model saw
The deal extract below has an owner name field that is blank on nearly every row and an owner id field that is populated. Using the owners table, resolve each open deal to an owner name. Report: number of deals resolved, deals whose owner id has no match in the owners table (list them), the owner ids that map to archived/deactivated owners (list their names), and the total pipeline amount per resolved owner. Do not guess a name for any unmatched id.
cost $0.1223 · 193s · in 44,373 / out 27,275 tok
✓ resolved_count✓ unmatched_none✓ deactivated_flagged✓ bryce_amount✓ alex_amount✓ no_invented

✓ Correct Answer (ground truth)

Plain-English answer: All 156 deals resolve to six named owners; jake_amount maps to Bryce Harmon ($1,054,144) and justin_amount to Alex Franklin ($624,310).
The answer in one line: jake_amount = 1,054,144; justin_amount = 624,310; resolved = 156
jake_amount1,054,144
justin_amount624,310
resolved156
deactivated
per_ownerBryce Harmon=1,054,144; Alex Franklin=624,310; Dana Mercer=341,195; Cole Ingram=288,161.43; Farid Osman=4,134; Elena Sinclair=2,100
graded checksresolved_count · unmatched_none · deactivated_flagged · bryce_amount · alex_amount · no_invented

Model's actual answer

show the model's full answer
Owner resolution for deals_open.csv (156 open deals) against owners.csv:

Note on the extract: the provided deals_open.csv contains no owner_name column at all (so there are no blank name cells to fill — the owner_id column is populated on all 156 rows). Every owner_id was resolved via owners.csv.

Resolved deals: 156 of 156 (100%).
Unmatched owner ids: none. No deal row carries an owner_id absent from owners.csv, so no deals are listed here.
Archived/deactivated owners appearing in the deals: none. The two archived owners in owners.csv — Gavin Porter (1520255671) and Hugo Lindqvist (77260721) — are referenced by zero deals. All six owner_ids present in deals map to active owners.

Pipeline per resolved owner:

| Owner (id) | Deals | Pipeline amount |
|---|---|---|
| Bryce Harmon (119337721) | 35 | 1,054,144.00 |
| Alex Franklin (84342457) | 67 | 624,310.00 |
| Dana Mercer (83155923) | 24 | 341,195.00 |
| Cole Ingram (83155924) | 22 | 288,161.43 |
| Farid Osman (716654662) | 7 | 4,134.00 |
| Elena Sinclair (701163055) | 1 | 2,100.00 |

Arithmetic checks:
- Deal count: 35 + 67 + 24 + 22 + 7 + 1 = 156 (matches row count).
- Total pipeline: 1,054,144.00 + 624,310.00 + 341,195.00 + 288,161.43 + 4,134.00 + 2,100.00 = 2,314,044.43, equal to the sum of all 156 row amounts.
- Spot check, Farid Osman (7 deals): 1,249 + 2,880 + 1 + 1 + 1 + 1 + 1 = 4,134.00 (Deal-499BF6, Deal-8BA24E, Deal-03BA69, Deal-117863, Deal-F17780, Deal-2B39B0, Deal-8FDCD2).
- Spot check, Elena Sinclair (1 deal): Deal-57FF13 = 2,100.00.

Caveats from the data as given: amounts of 1 (e.g., Deal-012CB1, Deal-483B2D, Deal-3795AD, and six Farid Osman DS1 rows) appear to be placeholder values and are included as-is; they inflate deal counts more than value (Farid Osman's 4,134.00 includes 5 of these $1 rows).

⚠️ File-mutation verifier: 2 file edit(s) FAILED this turn despite any wording above that may suggest otherwise. Run `git status` or `read_file` to confirm what actually landed.
  • `/Users/amaniphipps/.hermes/cache/scratch/deals_open.csv` — [write_file] Refusing to overwrite `/Users/amaniphipps/.hermes/cache/scratch/deals_open.csv`: `/Users/amaniphipps/.hermes/cache/scratch/deals_open.csv` exists but this task has not seen its full c…
  • `/Users/amaniphipps/.hermes/cache/scratch/owners.csv` — [write_file] Refusing to overwrite `/Users/amaniphipps/.hermes/cache/scratch/owners.csv`: `/Users/amaniphipps/.hermes/cache/scratch/owners.csv` exists but this task has not seen its full current c…
gong-hubspot-join-integrityPASS 1.00
Plain-English question: How many call records point at deal ids that do not exist in the open pipeline?
Prompt — what the model saw
Check the integrity of the call-to-deal mapping below. The gong table lists calls in the last 90 days keyed to HubSpot deal ids; deals_open lists the current open default-pipeline deals. Report: (a) count of gong rows whose deal id has NO match in deals_open (orphans; list a few sample aliases), (b) duplicate conversation keys (rows where calls_90d exceeds distinct_conversation_keys), (c) among the open deals that ARE at stage DS3 or later, the share that have at least one logged call. Give sample aliases for the orphan finding, all taken from the data. Note: gong_owner_id is a Gong user id, not a HubSpot owner id.
cost $0.1717 · 298s · in 52,867 / out 43,089 tok
✓ orphan_count✓ dup_keys_0✓ ds3_coverage✓ sample_ids_present

✓ Correct Answer (ground truth)

Plain-English answer: 30 Gong call rows are orphans keyed to dead deal ids, none are duplicates, and DS3 call coverage is 29.4%.
The answer in one line: orphans = 30; ds3_cov_pct = 29.4; dup = 0
orphans30
ds3_cov_pct29.4
dup0
orphan_aliasesDeal-8FA85D, Deal-8FC3F9, Deal-3B7945, Deal-42B265, Deal-9CCC42, Deal-36EA09
graded checksorphan_count · dup_keys_0 · ds3_coverage · sample_ids_present

Model's actual answer

show the model's full answer
Call-to-deal integrity check, gong_calls_by_deal_90d.csv (67 rows) vs deals_open.csv (156 rows, all deal_ids distinct):

(a) Orphaned gong rows: 30 of 67 (44.8%)
30 gong rows have an hs_deal_id with no match in deals_open (67 total - 30 orphans = 37 matched). Sample orphan aliases: Deal-8FA85D (46 calls), Deal-8FC3F9 (24), Deal-3B7945 (21), Deal-42B265 (21), Deal-9CCC42 (17), Deal-36EA09 (17), Deal-9A43B4 (14), Deal-605F3C (12), Deal-E2D34B (11), Deal-76821A (11).
Pattern: 25 of the 30 orphans have calls_with_ds2_date = 0 and calls_with_sqo_date = 0, consistent with calls keyed to deals that are not in the open default pipeline (closed, other pipeline, or deleted). The other 5 orphans - Deal-3F86A0, Deal-422BA6, Deal-B038F0, Deal-AC944F, Deal-38CA53 - have nonzero DS2/SQO dates on their calls but still no match in deals_open, so they likely reference deals that closed or moved out of the open set after the call data was pulled.

(b) Duplicate conversation keys: 0 rows
No row has calls_90d > distinct_conversation_keys. In every row the two are equal (e.g., Deal-C6FE92: 11 = 11; Deal-8FA85D: 46 = 46), so no duplicate conversation keys are detectable at the deal level.

(c) Call coverage among open deals at DS3 or later: 25/85 = 29.4%
- Open deals DS3+: 85 (DS3 = 61, DS4 = 14, DS5 = 10; 61+14+10 = 85)
- Of those, 25 have at least one gong call (matched by deal id; alias pairs agree in all 37 matched rows, e.g., Deal-C6FE92 DS3, Deal-25F752 DS4, Deal-D348E1 DS5)
- Share = 25 / 85 = 0.294, i.e. 29.4%; 60 of 85 (70.6%) late-stage open deals have no logged call in the 90-day window (examples without calls: Deal-9AAE5F, Deal-403845, Deal-6787C2, Deal-FC22A3, Deal-B936FE).

Consistency check: of the 37 matched gong rows, 25 map to DS3+ deals, 7 to DS2, 5 to DS1; the remaining 30 rows are the orphans above. Note as stated, gong_owner_id is a Gong user id and was not used for matching (matched on hs_deal_id = deal_id only).
snowflake-sql-trapped-catalogPASS 1.00
Plain-English question: Write one Snowflake query for first-month usage per customer using only the provided, trap-laden catalog.
Prompt — what the model saw
Using ONLY the catalog provided, write one Snowflake SQL query returning, per customer company, for its first calendar month as a customer: unique givers, recognition count, successful redemption count. Read the catalog notes carefully: some tables are stale or unpopulated and must be avoided, and one documented filter (excluding deleted givers) must NOT be applied to historical giving counts. Put a comment above each FROM or JOIN explaining the table choice. Return only fenced SQL.
cost $0.0453 · 107s · in 20,894 / out 10,926 tok
✓ avoids_stale_deals✓ avoids_unpopulated_live✓ uses_cohort_summary✓ succeeded_filter✓ no_giver_deletion_filter✓ returns_sql

✓ Correct Answer (ground truth)

Plain-English answer: A single correct SQL query that uses the right catalog tables and avoids the trap columns — graded on the exact tables and columns.
The answer in one line: use_table = PRODUCTION.PLG.COMPANY_COHORT_SUMMARY
avoid_stalePRODUCTION.HUBSPOT.DEALS (stale)
avoid_unpopulatedV2_LIVE.OBJECTS_DEALS (unpopulated)
forbidden_filterdeleted-giver exclusion must NOT be applied
outputone fenced SELECT with per-FROM/JOIN table-choice comment
required_filtersucceeded redemptions only
use_tablePRODUCTION.PLG.COMPANY_COHORT_SUMMARY
graded checksavoids_stale_deals · avoids_unpopulated_live · uses_cohort_summary · succeeded_filter · no_giver_deletion_filter · returns_sql

Model's actual answer

show the model's full answer
```sql
-- Per customer company, first calendar month as a customer: unique givers, recognition count,
-- successful redemption count.
--
-- Catalog constraints applied:
-- * PRODUCTION.PLG.COMPANY_COHORT_SUMMARY is the only catalog table with per-company, month-1 giving
--   metrics (M1_USERS, M1_GIVES, M1_REDEMPTIONS). No recognition/gives events table or givers table is
--   documented in the excerpt, so giving counts cannot be recomputed from events; the precomputed M1_*
--   columns are the only permitted source. M1_USERS is the only documented month-1 user-count metric and
--   is therefore used for unique givers (the catalog defines it no further).
-- * The "exclude deleted givers" filter is intentionally NOT applied: per business rule it understates
--   historical giving, so historical giving counts (M1_GIVES / M1_USERS) are taken unfiltered.
-- * PRODUCTION.DEPRECATED_RECOGNITION.REDEMPTION_RECORDS_V2 is the documented event-level source for
--   redemption counts and only STATE = 'succeeded' rows count, but the excerpt documents no company key or
--   timestamp column for it, so it cannot be joined per company / per month without inventing columns.
--   The first-month redemption count is therefore taken from M1_REDEMPTIONS (the cohort's documented
--   month-1 redemption metric).
-- * Stale / unpopulated / retired tables (PRODUCTION.HUBSPOT.DEALS,
--   HUBSPOT_HUB_1973303.V2_LIVE.OBJECTS_DEALS, PRODUCTION.HUBSPOT.GONG_HUBSPOT_MAP_FAST) are not used.
-- * The excerpt documents no company identifier column on COMPANY_COHORT_SUMMARY (one row per self-serve
--   company) and no anchor date for M1_* (SIGNUP_DATE vs ACTIVATED_DATE vs FIRST_SUB_PAYMENT_DATE), so no
--   company name is selected and no anchor filter is applied; each output row is one company's first month.

SELECT
    COHORT.M1_USERS       AS unique_givers,
    COHORT.M1_GIVES       AS recognition_count,
    COHORT.M1_REDEMPTIONS AS successful_redemption_count
-- COMPANY_COHORT_SUMMARY: the only per-company, first-calendar-month (M1) source in the catalog. One row
-- per self-serve company, so the M1_* columns already represent the first month as a customer and no join
-- or aggregation is needed. Stale deals tables and the retired GONG_HUBSPOT_MAP_FAST are irrelevant here.
FROM PRODUCTION.PLG.COMPANY_COHORT_SUMMARY AS COHORT
;
```
crm-completeness-enrichmentPASS 1.00
Plain-English question: Audit the CRM for missing and invalid fields, and propose the fixes.
Prompt — what the model saw
Audit the CRM extract for completeness and propose fixes. Deals need owner, stage, amount, close date, why-buys; companies need industry, employee count, HQ country; contacts need email, title, persona. Report completeness percent per field. Identify duplicate company clusters (name variants, shared domain) with a survivor per cluster. List invalid emails and domain mismatches. Fill missing company fields only where the enrichment export has a matching row; where CRM and enrichment disagree list both and recommend a source. Never invent a value. End with the 10 fixes with the most pipeline amount at stake.
cost $0.0609 · 116s · in 10,108 / out 17,455 tok
✓ dup_clusters✓ invalid_emails✓ survivor_language✓ completeness_pct✓ no_invent

✓ Correct Answer (ground truth)

Plain-English answer: 9 companies missing employee count, 6 missing HQ country, 4 invalid emails, 1 domain mismatch, 2 duplicate clusters, and 25 enrichment matches.
The answer in one line: n_contacts = 52; n_companies = 34; enrichment_matches = 25; invalid_email_count = 4; n_dup_clusters = 2; domain_mismatch_count = 1 (+4 more below)
n_contacts52
domain_mismatch_count1
enrichment_matches25
invalid_email_count4
n_companies34
n_dup_clusters2
company_missingindustry=0; employee_count=9; hq_country=6
domain_mismatchesCT-0011
duplicate_clusters['C-0A092931', 'C-0A092932'], ['C-0A092933', 'C-0A092934']
invalid_emailsCT-0010, CT-0080, CT-0081, CT-0192
graded checksdup_clusters · invalid_emails · survivor_language · completeness_pct · no_invent

Model's actual answer

show the model's full answer
Audit complete. One blocking gap up front: **no deals file was provided** (only companies.csv, contacts.csv, zoominfo_enrichment.csv). Therefore deal fields (owner, stage, amount, close date, why-buys) cannot be audited, and no pipeline amount can be attached to any fix. The final ranking is therefore by record/severity impact, with amounts explicitly unavailable.

## 1. Field completeness

Companies (34 rows):
- industry: 34/34 = 100.0% populated
- employee_count: 25/34 = 73.5% (9 missing: C-EC3025, C-96039F, C-44EA29, C-D04904, C-B23205, C-60C75F, C-7BBDFA, C-50D386, C-93C8BF)
- hq_country: 28/34 = 82.4% (6 missing: C-2D1F1B, C-D73B89, C-44EA29, C-D04904, C-2C60E5, C-EE9FFB)

Contacts (52 rows):
- email: 52/52 = 100.0% populated, but 48/52 = 92.3% valid (see §3)
- title: 39/52 = 75.0% (13 missing: CT-0000, CT-0022, CT-0072, CT-0080, CT-0081, CT-0092, CT-0120, CT-0121, CT-0122, CT-0132, CT-0141, CT-0162, CT-0170)
- persona: 37/52 = 71.2% (15 missing: CT-0000, CT-0022, CT-0041, CT-0060, CT-0070, CT-0081, CT-0082, CT-0092, CT-0110, CT-0132, CT-0162, CT-0171, CT-0172, CT-0180, CT-0181)

Deals: no file supplied — completeness not computable.

Coverage note: 14 companies have zero contacts (C-2C60E5, C-B97B4E, C-2D7423, C-7BBDFA, C-50D386, C-BA969B, C-332637, C-93C8BF, C-EE9FFB, C-C9BB20, C-0A092931, C-0A092932, C-0A092933, C-0A092934).

## 2. Duplicate company clusters

Name-variant clustering is not possible: company_alias values are opaque hashes with no name field. Shared-domain clusters found (2 clusters, 4 rows):

Cluster A — acme-corp.com: C-0A092931 (Technology, 500, US) vs C-0A092932 (tech, 510, USA).
Survivor: C-0A092931 — canonical industry/country values already conform; merge C-0A092932 into it. Open conflict: headcount 500 vs 510; no ZoomInfo row for acme-corp.com, so no arbiter; keep both values flagged pending review.

Cluster B — globex.io: C-0A092933 (SaaS, 200, US) vs C-0A092934 (Technology, 200, US).
Survivor: C-0A092934 — industry "Technology" matches the CRM canonical vocabulary; employee_count and hq_country identical, so merge is otherwise lossless. Open conflict: "SaaS" vs "Technology" — resolve to the CRM's controlled list.

No other shared domains (all other 30 domains unique).

## 3. Email validity and domain match

Invalid emails (4, local-part@ with no domain): CT-0010 (user0@), CT-0080 (user0@), CT-0081 (user1@), CT-0192 (user2@).
Domain mismatch (1): CT-0011 user1@other-domain.com vs record domain 66d1fc.com — either a personal-address entry or a wrong company link; verify before enriching.

## 4. Enrichment fills and CRM/ZI conflicts

ZoomInfo covers 25 of 34 company domains. 9 companies have no ZI row: C-BA969B, C-332637, C-93C8BF, C-EE9FFB, C-C9BB20, C-0A092931, C-0A092932, C-0A092933, C-0A092934.

Fillable employee_count from ZI (8 companies): C-EC3025, C-96039F, C-44EA29, C-D04904, C-B23205, C-60C75F, C-7BBDFA, C-50D386 — all ZI values = 400. C-93C8BF not fillable (no ZI row). This lifts employee_count completeness to 33/34 = 97.1% (25+8=33; 33/34 = 97.06%).
Fillable hq_country from ZI: 0. Every company with missing hq_country has a blank ZI country (C-2D1F1B, C-D73B89, C-44EA29, C-D04904, C-2C60E5) or no ZI row (C-EE9FFB). Do not invent.

Conflicts where both sources have values:
- employee_count: no true conflicts among rows where both present.
- hq_country: 10 format-only conflicts (same country, different spelling): C-66D1FC, C-950043, C-77A95A, C-B23205, C-D0662E = "US" vs ZI "United States"; C-EC3025, C-96039F, C-E51FB7, C-425E2A, C-2D7423 = "USA" vs ZI "United States". CRM uses three US spellings (US 9, USA 6, United States 2). Recommendation: keep CRM values but normalize to one convention (ZI full name "United States" or ISO "US"); this is a formatting fix, not a data conflict.
- industry: CRM "tech"/"Tech " (5 rows: C-66D1FC, C-44EA29, C-60C75F, C-425E2A, C-BA969B, C-93C8BF, C-C9BB20, C-0A092932) maps to ZI "Computer Software" — same broad segment, different taxonomy. CRM "health care" (C-7BBDFA, C-50D386) matches ZI's non-canonical "health care". Recommendation: adopt one controlled vocabulary (CRM "Technology"/"Healthcare") and map ZI's "Computer Software"/"health care" into it; do not overwrite CRM categories wholesale. 11 rows total need normalization (8 tech variants, 2 health care, 1 SaaS).
- Headcount sanity note: ZI returns 400 for all 8 fillable rows; plausible but flag as possible ZI default before trusting as ground truth.

## 5. Top 10 fixes — pipeline amount at stake: UNAVAILABLE

No deals/amount data exists in the extract, so no fix can be dollar-ranked. Ranked instead by records affected and downstream risk:

1. Merge acme-corp.com cluster (2 companies) — survivor C-0A092931; duplicate records will double-count pipeline once deals exist.
2. Merge globex.io cluster (2 companies) — survivor C-0A092934.
3. Fix 4 invalid emails (CT-0010, CT-0080, CT-0081, CT-0192) — these contacts are unreachable for engagement/attribution.
4. Resolve CT-0011 domain mismatch (user1@other-domain.com vs 66d1fc.com) — wrong-domain contacts poison account mapping.
5. Apply 8 ZI employee_count fills (all = 400) — lifts company headcount completeness 73.5% → 97.1%.
6. Normalize hq_country across 28 populated rows (3 US spellings) and flag the 6 blanks for a non-ZI source (ZI is also blank — do not invent).
7. Normalize industry across 11 rows (tech/Tech/health care/SaaS → canonical set).
8. Backfill 13 missing contact titles.
9. Backfill 15 missing contact personas — blocks buyer-role coverage checks on deals.
10. Enrich/attach contacts for the 14 companies with zero contacts and 9 companies with no ZI row (both duplicate clusters and C-EE9FFB, C-BA969B, C-332637, C-C9BB20, C-93C8BF included).

Required to complete this audit as specified: a deals extract with owner, stage, amount, close date, why-buys, and company links — without it, deal completeness and pipeline-dollar prioritization remain impossible.
deal-intelligence 5 tests
closed-lost-classificationPASS 1.00
Plain-English question: Why did we lose the last six months of deals — one primary category per deal, and whose side is the reason on?
Prompt — what the model saw
Classify every deal lost in the last 6 months below into one primary category: pricing, competitor, no decision, timing, product gap, champion left, or other. Also assign a side (Bonusly, buyer, unknown). Then summarize: category counts, the side split, the number of deals where the structured closed_lost_tag clearly disagrees with the free-text reason, and the two patterns most worth acting on. Base everything only on the tag and free-text provided.
cost $0.1404 · 380s · in 10,301 / out 43,858 tok
✓ total_90✓ competitor_top✓ category_counts_present✓ side_split✓ two_patterns

✓ Correct Answer (ground truth)

Plain-English answer: Every closed-lost deal gets exactly one category (pricing, competitor, no decision, timing, product gap, champion left, other) plus a side, per the computed classification.
The answer in one line: total_lost_6mo = 90; top_reason = competitor (26)
total_lost_6mo90
categoriespricing, competitor, no decision, timing, product gap, champion left, other
patternstwo most actionable, tag-vs-text disagreements counted
sideBonusly / buyer / unknown per deal
top_reasoncompetitor (26)
graded checkstotal_90 · competitor_top · category_counts_present · side_split · two_patterns

Model's actual answer

show the model's full answer
Per-deal classification (file order, deal_alias as given; * = tag/text disagreement):

```
alias     category      side
DB0AAC    timing        buyer
F7F635    competitor    unknown
AC944F    no decision   buyer
214060    no decision   buyer
91A056    timing        buyer
29326C    timing        buyer
5DB9B0    other         unknown
831B7B    timing        buyer
F97C37    competitor    Bonusly
13E9CF    no decision   buyer
39E25C    timing        buyer
7ED004    pricing       buyer
21B045    no decision   buyer
B3ABED    timing        buyer
422BA6    competitor    Bonusly
ED9AE7    no decision   buyer
988493    no decision   buyer
381C8C    competitor    unknown
F308CA    no decision   buyer
F1E8A6    competitor    unknown
B6AC09    timing        buyer
70F704    no decision   buyer
E6E80A    timing        buyer
B038F0    timing        buyer
4664E1    no decision   buyer
175756    timing        buyer
E74A73    no decision   buyer
DDAB52    competitor    Bonusly
ACE061    competitor    unknown
BB78F3    timing        buyer
D48E0B    no decision   buyer
15DA99    timing        buyer
F4AF5D    timing        buyer
79B7A1    timing        buyer
583ADB    no decision   buyer
8E27DA    no decision   buyer   *
2D2F8D    competitor    unknown
E0441F    no decision   buyer
7CB44D    no decision   buyer
0F96AA    competitor    unknown
1BCA50    pricing       buyer   *
7CC678    competitor    unknown
FAC17C    no decision   buyer
242273    competitor    Bonusly
50E5D8    no decision   buyer
A2C349    competitor    buyer
9F176A    timing        buyer
7B2236    pricing       buyer
AFA56C    no decision   buyer
C7156E    competitor    unknown
C33D91    pricing       buyer
9048EB    product gap   Bonusly *
5E64CE    timing        buyer
8A0992    competitor    buyer
D0C698    competitor    buyer
69CF3D    timing        buyer
ECBF89    timing        buyer
3618CC    product gap   Bonusly
EECC02    competitor    unknown
5AD03E    product gap   Bonusly *
D1A623    timing        buyer
413C56    no decision   buyer
47F1A1    competitor    buyer
BF2A98    competitor    buyer
2A292B    no decision   buyer
D1AABF    no decision   buyer
FEDBCB    timing        buyer
1E7DA9    competitor    unknown
2BBA21    no decision   buyer
286F9C    competitor    Bonusly
7FBAC6    no decision   buyer
369281    competitor    buyer
386F6E    no decision   buyer
9FCD0D    competitor    buyer
55867E    no decision   unknown *
DAFB82    pricing       buyer
2FEDDB    no decision   buyer
64B19A    competitor    buyer
3F86A0    no decision   buyer
096750    no decision   buyer
F325A5    champion left buyer
ABD14C    no decision   buyer
79E61A    no decision   buyer
8A119B    pricing       buyer
AE7C4E    no decision   buyer
DAB4F1    no decision   buyer
B4B50F    no decision   buyer
981AD4    product gap   Bonusly
DC77FE    competitor    Bonusly
5885B9    no decision   buyer
```

Summary arithmetic (90 deals total):

Category counts: no decision 34, competitor 24, timing 20, pricing 6, product gap 4, champion left 1, other 1. Sum: 34+24+20+6+4+1+1 = 90.
Side split: buyer 68, unknown 12, Bonusly 10. Sum: 68+12+10 = 90.

Tag vs free-text disagreements: 5.
- Deal-8E27DA: tag Feature Request (implies product gap), text says "didn't want R&R, moved forward with just a swag provider" -> actually no decision.
- Deal-5AD03E: tag Competitor, text "Wanted more defined budget access" names no competitor; it's a budget-access feature ask -> product gap.
- Deal-9048EB: tag MIA, text says "bad fit based on their desired setup and multiple feature gaps" -> product gap, not unresponsiveness.
- Deal-1BCA50: tag Competitor, text says "It was mostly about the budget and details regarding gift cards" -> pricing is the stated primary, competitor was secondary.
- Deal-55867E: tag "Timing (1 year or more)", text is a polite final-sounding decline ("don't think we'll be moving forward... at this time") with no deferral or reconnect date -> no decision.

Note (excluded from the count, but a hygiene problem): 5 deals carry the opaque tag "Lost DM" (ED9AE7, 70F704, FAC17C, 3618CC, F325A5). The tag carries no category, so it cannot "clearly disagree"; the free text supplies the real reason in all five. Similarly, 10 of the 24 competitor-tagged deals (e.g. F7F635, 381C8C, F1E8A6, 0F96AA, C7156E, 1E7DA9, 7CC678, EECC02, 2D2F8D, ACE061) have no reason in the text at all ("went in another direction", "nothing specific provided"), which is why 12 deals land in "unknown".

Two patterns worth acting on:

1. Non-competitive stall dominates and is systematically unworked: no decision (34) + timing (20) = 54 of 90 (60%). The MIA cluster (22 deals) shows the same failure mode repeated: "ignored multiple pieces of outreach from me and the ADR" (F308CA, 4664E1, 7CB44D, 2BBA21, 096750) — reps burn multiple touches then close MIA with no structured exit or re-entry. It is also concentrated: owner 119337721 has 7 MIA closes and owner 83155924 has 7. Meanwhile many timing closes carry explicit future dates ("reconnect early 2027" 91A056, E6E80A, B038F0, 15DA99; "revisiting in 2027" B6AC09, 175756; budget "not until 2028" B3ABED, DAFB82; "contract through October 2027" 5E64CE). Action: a dated re-engagement system for the 2027/2028 pause cohort plus a MIA playbook (touch caps, qualification bar, auto re-entry dates) targeted at the two owners carrying half the MIA closes.

2. Competitor losses are half-blind and the explained half points at a short, concrete gap list: 10 of 24 competitor losses have no reason captured at all. Of the 14 explained, 8 are incumbent/preference-driven (A2C349 Awardco, 47F1A1 WorkTango, BF2A98 HiThrive, 64B19A Motivosity, D0C698 Kudos, 369281 paylocity, 9FCD0D and 8A0992 Canadian-provider preference) — largely buyer-side and hard to win. The 6 Bonusly-side losses (F97C37, 422BA6, DDAB52, 242273, 5AD03E, 286F9C, with DC77FE and 242273 detailing it) cluster on: breadth of offering beyond R&R (F97C37, DDAB52), a missing preferred-partner/integration motion (422BA6 lost to an ADP TotalSource PEO partner), customization of the points currency (DC77FE "label points as dollars", 242273 "digitize internal points currency... spend at onsite facilities"), and budget-access controls (5AD03E). Action: make loss-reason capture mandatory at closed-lost for competitor tags, and treat the points-currency customization + PEO-partnership gaps as product/strategy items with named deals behind them.
pipeline-tieringPASS 1.00
Plain-English question: Score all 156 open deals into exactly one of LOCK / ACTION / BUILD / REVIVE / WATCH / RISKY.
Prompt — what the model saw
Score the open pipeline below on these signals: forecast category, stage, meetings_30d, contact count, and engagement recency. Assign exactly one tier per deal: LOCK, ACTION, BUILD, REVIVE, WATCH, or RISKY (RISKY = the forecast category disagrees with the engagement evidence). Never tier a deal LOCK if it has zero meetings_30d. Do NOT list every deal. Return ONLY a compact JSON object with exactly these keys: - "tier_counts": object mapping each tier to the number of deals in it (must sum to the total deal count), - "tier_examples": object mapping each tier to up to 3 example deal aliases from the data, - "risky_deals": array of the deal aliases you tiered RISKY, - "lock_violations": number of deals tiered LOCK that have zero meetings_30d (should be 0), - "pipeline_shape": one short paragraph on the overall pipeline shape. Note: inbound_emails_30d is 0 for every row due to a data defect, so treat meetings_30d as the inbound signal.
cost $0.2408 · 295s · in 77,805 / out 59,378 tok
✓ json_returned✓ tier_vocabulary✓ tier_counts_sum✓ risky_deals_listed✓ no_lock_violation_claimed

✓ Correct Answer (ground truth)

Plain-English answer: All 156 open deals tiered exactly once across the six tiers using forecast category, stage, meetings, contacts, and recency.
The answer in one line: total = 156
total156
graded checksjson_returned · tier_vocabulary · tier_counts_sum · risky_deals_listed · no_lock_violation_claimed

Model's actual answer

show the model's full answer
```json
{
  "tier_counts": {
    "LOCK": 3,
    "ACTION": 21,
    "BUILD": 39,
    "REVIVE": 27,
    "WATCH": 59,
    "RISKY": 7
  },
  "tier_examples": {
    "LOCK": ["Deal-D348E1", "Deal-C26D20", "Deal-403845"],
    "ACTION": ["Deal-25F752", "Deal-E53952", "Deal-5408B0"],
    "BUILD": ["Deal-C6FE92", "Deal-D73B89", "Deal-523604"],
    "REVIVE": ["Deal-2D1F1B", "Deal-66D1FC", "Deal-950043"],
    "WATCH": ["Deal-6787C2", "Deal-012CB1", "Deal-C9C286"],
    "RISKY": ["Deal-BA571A", "Deal-FC22A3", "Deal-7BBDFA"]
  },
  "risky_deals": [
    "Deal-BA571A",
    "Deal-FC22A3",
    "Deal-7BBDFA",
    "Deal-4A13AD",
    "Deal-15D24F",
    "Deal-635B8E",
    "Deal-ED725A"
  ],
  "lock_violations": 0,
  "pipeline_shape": "156 open deals totaling ~$2.31M, and 79% of that value ($1.83M) sits in PIPELINE forecast category, so the book is top-heavy and early: 66% of deals are DS1-DS2 and only 11 are COMMIT. Only 3 deals earn LOCK (Deal-D348E1, Deal-C26D20, Deal-403845: commit stage plus 1-4 meetings in the last 30 days), worth $36,270 combined, about 1.6% of pipeline value; the rest of the commit column (e.g. Deal-547B2B, Deal-B7EBD1, Deal-A2B47C) shows email activity but zero 30-day meetings, so it lands in ACTION rather than LOCK. Engagement is thin across the board: 101 of 156 deals (65%) have zero meetings_30d, which is why BUILD (meeting momentum at early/mid stages) is 39, REVIVE (prior engagement history but no calls/meetings and stale or email-only 30-day activity) is 27, and WATCH (weak, new, or missing engagement, including the 8 placeholder $1 deals and the two deals with no engagement row at all, Deal-3EED2C and Deal-57FF13) is 59. The 7 RISKY deals are all BEST_CASE at DS3-DS4 (Deal-BA571A, Deal-FC22A3, Deal-7BBDFA, Deal-4A13AD, Deal-15D24F, Deal-635B8E, Deal-ED725A): forecast says near-close, but zero 30-day meetings and calls, thin 90-day engagement (5-20 touches), and recency of 14-58 days say the deals are not being worked. Bottom line: a large unworked top, a handful of genuinely live mid-funnel deals, and a commit column that is almost entirely unvalidated by meeting activity."
}
```
call-transcript-extractionPASS 1.00
Plain-English question: Extract the CRM write-back fields from each call transcript as JSON.
Prompt — what the model saw
For each transcript, extract CRM write-back fields as JSON: why-buys (prospect statements only), pain points, stakeholders from the speaker list, budget signal (prospect-stated or null), timeline signal, competitor mentioned (only if the prospect raised it), next step (only if explicitly agreed), objections, confidence. Never fill a field from what the rep said or hoped.
cost $0.0245 · 61s · in 2,786 / out 7,236 tok
✓ json_emitted✓ budget_signal✓ null_discipline✓ stakeholders✓ competitor_prospect_only

✓ Correct Answer (ground truth)

Plain-English answer: Per transcript: prospect-only why-buys, pain points, stakeholders, budget and timeline signals, and competitors — e.g. TX-001: '$40k earmarked', live before open enrollment.
The answer in one line: rep = Alex Franklin
allowed_newTX-001, TX-002, TX-003, TX-004, TX-005, TX-006, Achievers, Bucketlist, Nectar
per_transcript{"TX-001": {"why_buys": ["automating anniversary and birthday awards"], "pain_points": ["HR team of three cannot keep up with it manually", "track everything in a spreadsheet, and people slip through
repAlex Franklin
rep_sourced_nulls{"TX-003": ["budget_signal \u2014 only the rep mentioned pricing/budget"], "TX-004": ["next_step \u2014 rep proposed, prospect said 'no promises'"], "TX-006": ["budget_signal \u2014 rep decoy, prospec
graded checksjson_emitted · budget_signal · null_discipline · stakeholders · competitor_prospect_only

Model's actual answer

show the model's full answer
[
  {
    "transcript_id": "TX-001",
    "deal_alias": "Deal-CFE7F4",
    "why_buys": [
      "Automating anniversary and birthday awards would be the big win; HR team of three cannot keep up with it manually (Prospect (VP People))"
    ],
    "pain_points": [
      "Everything is tracked in a spreadsheet and people slip through the cracks (Prospect (HR Admin))",
      "HR team of three cannot keep up with manual anniversary/birthday awards (Prospect (VP People))"
    ],
    "stakeholders": [
      {"speaker": "Alex Franklin", "side": "vendor"},
      {"speaker": "Prospect (VP People)", "side": "prospect"},
      {"speaker": "Prospect (HR Admin)", "side": "prospect"}
    ],
    "budget_signal": "$40k earmarked for engagement tools this fiscal year (Prospect (VP People))",
    "timeline_signal": "Would have this live before open enrollment in November (Prospect (VP People))",
    "competitor_mentioned": "Achievers — looked at last year, 'too heavy for a team our size' (Prospect (VP People))",
    "next_step": "Security review on September 12 (explicitly agreed by Prospect (VP People))",
    "objections": [
      "Need SSO and audit logs for IT to sign off (Prospect (HR Admin))"
    ],
    "confidence": {"level": "high", "reason": "Prospect-stated budget, timeline, competitor, objection, and an explicitly agreed next step."}
  },
  {
    "transcript_id": "TX-002",
    "deal_alias": "Deal-70BB30",
    "why_buys": [
      "Tie recognition to retention for the hourly workforce (Prospect (Head of Total Rewards))"
    ],
    "pain_points": [
      "Regretted turnover in the hourly workforce is over 30% (Prospect (Head of Total Rewards))"
    ],
    "stakeholders": [
      {"speaker": "Alex Franklin", "side": "vendor"},
      {"speaker": "Prospect (Head of Total Rewards)", "side": "prospect"},
      {"speaker": "Prospect (CFO)", "side": "prospect"}
    ],
    "budget_signal": "$25k pilot budget approved by Finance for this quarter (Prospect (CFO))",
    "timeline_signal": "Decision wanted by end of September (Prospect (CFO))",
    "competitor_mentioned": null,
    "next_step": "Send the pilot agreement; CFO will route it to legal this week (explicitly agreed by Prospect (CFO))",
    "objections": [
      "Integration with Workday has to be rock solid — CFO's one condition (Prospect (CFO))"
    ],
    "confidence": {"level": "high", "reason": "No competitor raised ('first vendor we've had a real demo with'); all other fields prospect-stated and explicit."}
  },
  {
    "transcript_id": "TX-003",
    "deal_alias": "Deal-530B50",
    "why_buys": [
      "Make recognition visible across the 12 retail locations (Prospect (People Ops Manager))"
    ],
    "pain_points": [
      "Store managers have zero budget autonomy for on-the-spot recognition today (Prospect (People Ops Manager))"
    ],
    "stakeholders": [
      {"speaker": "Alex Franklin", "side": "vendor"},
      {"speaker": "Prospect (People Ops Manager)", "side": "prospect"}
    ],
    "budget_signal": null,
    "timeline_signal": "No rush on our side until Q1 (Prospect (People Ops Manager))",
    "competitor_mentioned": "Bucketlist — CEO used it at her last company and liked it (Prospect (People Ops Manager))",
    "next_step": "Schedule a call with the CEO; prospect will send two times (explicitly agreed by Prospect (People Ops Manager))",
    "objections": [
      "The CEO has to be sold first — she decides anything people-related (Prospect (People Ops Manager))",
      "No urgency: 'no rush on our side until Q1' (Prospect (People Ops Manager))"
    ],
    "confidence": {"level": "medium", "reason": "Single prospect speaker; no prospect-stated budget (the $8/employee/month figure is rep-only and excluded); timeline is a deferral."},
    "notes": "budget_signal is null because the only price mention ($8 per employee per month) was made by the rep, Alex Franklin."
  },
  {
    "transcript_id": "TX-004",
    "deal_alias": "Deal-180D02",
    "why_buys": [
      "Consolidate three separate recognition tools into one (Prospect (VP People))"
    ],
    "pain_points": [
      "Paying for three tools and none of them talk to the HRIS (Prospect (VP People))"
    ],
    "stakeholders": [
      {"speaker": "Alex Franklin", "side": "vendor"},
      {"speaker": "Prospect (VP People)", "side": "prospect"},
      {"speaker": "Prospect (IT Security Lead)", "side": "prospect"}
    ],
    "budget_signal": "Under $15k annually can be approved without going to the board (Prospect (VP People))",
    "timeline_signal": "Procurement cycle runs six to eight weeks minimum (Prospect (IT Security Lead))",
    "competitor_mentioned": null,
    "next_step": null,
    "objections": [
      "Security review took three months for the last vendor — stated hesitation (Prospect (IT Security Lead))"
    ],
    "confidence": {"level": "high", "reason": "Budget authority, procurement timeline, and objection all prospect-stated."},
    "notes": "next_step is null: the rep asked to lock a CFO follow-up, but the prospect replied 'Maybe — I need to check her calendar, no promises' (no explicit agreement). No competitor was raised by the prospect."
  },
  {
    "transcript_id": "TX-005",
    "deal_alias": "Deal-F8767A",
    "why_buys": [
      "Automate service milestones (Prospect (HR Director))",
      "Analytics on recognition equity across departments (Prospect (HR Director))"
    ],
    "pain_points": [
      "Night-shift teams feel invisible — engagement scores run 20 points lower (Prospect (People Ops Coordinator))"
    ],
    "stakeholders": [
      {"speaker": "Alex Franklin", "side": "vendor"},
      {"speaker": "Prospect (HR Director)", "side": "prospect"},
      {"speaker": "Prospect (People Ops Coordinator)", "side": "prospect"}
    ],
    "budget_signal": "$12k approved under the engagement line (Prospect (HR Director))",
    "timeline_signal": "Needs to be running before the January all-hands (Prospect (HR Director))",
    "competitor_mentioned": "Nectar — mid-pilot with them now, 'you'd need to beat that experience' (Prospect (HR Director))",
    "next_step": "Present to the exec team on October 2 (explicitly agreed by Prospect (HR Director))",
    "objections": [
      "Exec team is skeptical after a failed rollout two years ago (Prospect (HR Director))",
      "Must beat the current Nectar pilot experience (Prospect (HR Director))"
    ],
    "confidence": {"level": "high", "reason": "Budget, timeline, competitor, objections, and agreed next step all prospect-stated."}
  },
  {
    "transcript_id": "TX-006",
    "deal_alias": "Deal-EE195F",
    "why_buys": [
      "Cut the admin time on service awards (Prospect (HR Manager))"
    ],
    "pain_points": [
      "HR Manager personally spends five hours a month ordering and shipping plaques (Prospect (HR Manager))"
    ],
    "stakeholders": [
      {"speaker": "Alex Franklin", "side": "vendor"},
      {"speaker": "Prospect (HR Manager)", "side": "prospect"}
    ],
    "budget_signal": null,
    "timeline_signal": "Q1 start is realistic; fiscal year resets in February (Prospect (HR Manager))",
    "competitor_mentioned": null,
    "next_step": "Send the one-page overview; HR Manager will forward it to the COO this week (explicitly agreed by Prospect (HR Manager))",
    "objections": [
      "Comparing the purchase against just doing it internally; no other vendor (Prospect (HR Manager))",
      "COO usually prefers building things in-house (Prospect (HR Manager))"
    ],
    "confidence": {"level": "high", "reason": "Clear prospect statements; budget amount absent because prospect explicitly declined to give one ('Budget isn't the issue — time is')."},
    "notes": "budget_signal is null: the pricing flexibility offer was rep-only, and the prospect stated budget is not the issue without naming a number. competitor_mentioned is null: no vendor raised; the internal-build option is captured under objections."
  }
]

Summary of nulls and exclusions (per the rule that rep statements never fill fields):
- TX-003 / Deal-530B50: budget null (rep quoted $8/employee/month; prospect stated none).
- TX-004 / Deal-180D02: next_step null (prospect said "Maybe... no promises" to the CFO follow-up); competitor null (none raised).
- TX-002 / Deal-70BB30 and TX-006 / Deal-EE195F: competitor null (prospect confirmed no other vendors; TX-006's internal-build comparison recorded as an objection, not a competitor).
- TX-006 / Deal-EE195F: budget null (prospect declined to give a number).
next-to-close0.75
Plain-English question: Which 3 deals are closest to signature right now?
Prompt — what the model saw
Which 3 deals are closest to signature right now? Use the CRM data and the Slack thread; Slack often has the real status before the CRM does. For each: deal, amount, why it is close, what is left. Under 250 words.
cost $0.0115 · 31s · in 2,212 / out 3,021 tok
✗ exactly_three✓ top3_named✓ slack_first✓ blocked_excluded

✓ Correct Answer (ground truth)

Plain-English answer: Deal-547B2B ($11,200), Deal-B7EBD1 ($9,000), and Deal-403845 ($9,000) — all DS5 COMMIT deals closing September 10-11, with what is left on each.
blocked_excludeddeal_alias=Deal-2465CE; reason=champion left + procurement freeze (Slack #sales-sparring 2026-09-04)
exactly_three1
ranking_ruleSlack signature-imminent deal ranks first; remaining ranked by stage (DS5>DS4), forecast (COMMIT>BEST_CASE>PIPELINE), close_date asc, amount desc
top3rank=1; deal_alias=Deal-547B2B; amount=11200; stage=DS5; forecast_category=COMMIT; close_date=2026-09-11 | rank=2; deal_alias=Deal-B7EBD1; amount=9000; stage=DS5; forecast_category=COMMIT; close_date=2026-09-10 | rank=3; deal_alias=Deal-403845; amount=9000; stage=DS5; forecast_category=COMMIT; close_date=2026-09-11
graded checksexactly_three · top3_named · slack_first · blocked_excluded

Model's actual answer

show the model's full answer
Anchored to the Slack snapshot date, 2026-09-04.

1. Deal-547B2B — $11,200 — DS5/COMMIT, close 2026-09-11 (Alex Franklin)
Why close: redlines came back clean, signing page is out, VP People said they sign tomorrow. Only CRM-corroborated and Slack-confirmed signature-imminent deal.
What's left: the counter-signature. Nothing else named.

2. Deal-403845 — $9,000 — DS5/COMMIT, close 2026-09-11 (CRM owner Alex Franklin; Dana Mercer reporting on it)
Why close: order form is with the customer's finance team — past redlines and pricing, downstream of legal.
What's left: finance approval/PO, then counter-signature.

3. Deal-A2B47C — $6,360 — DS5/COMMIT, close 2026-09-11 (Alex Franklin)
Why close: DS5/COMMIT with a positive Slack update — "still warm, just normal legal-review pace," no blockers.
What's left: legal review; no order form or signature date yet, so the weakest of the three.

Sum: 11,200 + 9,000 + 6,360 = $26,560.

Excluded despite CRM DS5/COMMIT: Deal-2465CE ($5,400, close 9/10) — champion left, procurement froze new vendors, pulled from commit, now Q4. CRM is stale here; CRM alone would have ranked it #2 (11,200 + 9,000 + 5,400 = $25,600).

No Slack signal: Deal-B7EBD1 ($9,000, DS5/COMMIT, close 9/10, Dana Mercer) has the earliest close date but no thread update — swap candidate for #3 if verified. Deal-D348E1 ($13,770, DS5/COMMIT, close 10/15) is warm but 34 days out.
product-gap-detectionPASS 1.00
Plain-English question: Which prospect complaints are true product gaps versus rollout or enablement issues?
Prompt — what the model saw
Surface true product gaps raised by prospects in active deals. For each candidate: quote the prospect line, classify as real gap, plan gate, or rollout/enablement issue using the product docs excerpt, and name the deal and amount. Only real gaps go in the summary. Rep language does not count as prospect voice.
cost $0.0082 · 21s · in 1,482 / out 2,047 tok
✓ real_gaps✓ plan_gate_classified✓ enablement_classified✓ gap_summary

✓ Correct Answer (ground truth)

Plain-English answer: Two real gaps (SCIM provisioning, ADP Workforce Now integration) separated from plan gates and enablement issues, each with the prospect quote and deal.
allowed_newTG-001, TG-002, TG-003, TG-004, TG-005, SCIM user provisioning, ADP Workforce Now integration, Custom report builder
enablement_issuestranscript_id=TG-004; deal_alias=Deal-84DBA6; amount=16000; feature=Slack integration recognition adoption; quote=We turned on the Slack integration in July, but half our managers still haven't given a single recognition — nobody ever trained them on it.; why_enablement=Slack integration exists on all plans; managers were never trained
excluded_rep_voicetranscript_id=TG-005; line=Full transparency — we don't have a native mobile app for hourly workers yet, though it's on the roadmap.; reason=rep voice, not prospect-raised
plan_gatestranscript_id=TG-003; deal_alias=Deal-CFE7F4; amount=32000; feature=Custom report builder; tier_required=Enterprise; quote=I tried to build a custom report combining recognition frequency with turnover and couldn't — is that possible on our current plan?
real_gapstranscript_id=TG-001; deal_alias=Deal-EC3025; amount=62000; feature=SCIM user provisioning; quote=We need SCIM user provisioning — our security team will not approve any tool that requires manual account management.; why_real=not listed in product docs at any tier | transcript_id=TG-002; deal_alias=Deal-D0D6B5; amount=23800; feature=ADP Workforce Now integration; quote=Our payroll runs on ADP Workforce Now and I don't see ADP anywhere in your integration list — that's a dealbreaker for us.; why_real=supported HRIS list is Workday/BambooHR/Gusto/Rippling only
graded checksreal_gaps · plan_gate_classified · enablement_classified · gap_summary

Model's actual answer

show the model's full answer
Candidate review (5 transcripts; 4 contain prospect-raised items; 1 excluded as rep-only). Deal amounts are not present in either provided file (transcripts_gaps.csv carries no amount column; product_docs.md has none), so amounts cannot be cited and are noted as missing rather than invented.

1. TG-001 — Deal-EC3025 (amount: not in provided data)
Prospect quote: "We need SCIM user provisioning — our security team will not approve any tool that requires manual account management."
Classification: Real gap. The docs state SCIM user provisioning is NOT currently listed as a supported capability, on any plan. The security team's approval condition cannot be met today.

2. TG-002 — Deal-D0D6B5 (amount: not in provided data)
Prospect quote: "Our payroll runs on ADP Workforce Now and I don't see ADP anywhere in your integration list — that's a dealbreaker for us."
Classification: Real gap. Docs list HRIS integrations as Workday, BambooHR, Gusto, Rippling only, and explicitly state ADP Workforce Now is NOT currently listed as supported. The stated dealbreaker is confirmed by the docs.

3. TG-003 — Deal-CFE7F4 (amount: not in provided data)
Prospect quote: "I tried to build a custom report combining recognition frequency with turnover and couldn't — is that possible on our current plan?"
Classification: Plan gate. The custom report builder exists but is Enterprise-only per the tier table; the prospect's current plan excludes it. Capability gap is tier-based, not product-missing.

4. TG-004 — Deal-84DBA6 (amount: not in provided data)
Prospect quote: "We turned on the Slack integration in July, but half our managers still haven't given a single recognition — nobody ever trained them on it."
Classification: Rollout/enablement issue. Slack integration is supported on all plans and is already enabled; the problem is missing manager training/adoption, not a missing feature.

5. TG-005 — Deal-36C33F (amount: not in provided data)
Prospect quote: "Good to know. The web version should be fine for our office staff for now."
Classification: Not a gap (excluded). The missing native mobile app was raised by the rep (Alex Franklin), not the prospect; rep language does not count as prospect voice. The prospect explicitly accepted the web version, so no prospect-raised gap exists.

Summary — true product gaps only:
- SCIM user provisioning absent — Deal-EC3025 (amount not in provided data)
- ADP Workforce Now integration absent — Deal-D0D6B5 (amount not in provided data)

Tally: 5 transcripts reviewed = 2 real gaps + 1 plan gate + 1 rollout/enablement + 1 excluded rep-raised item.
rep-performance 5 tests
stale-pipeline-by-repPASS 1.00
Plain-English question: Which open deals have had no email, call, or meeting in the last 7 days?
Prompt — what the model saw
The snapshot date is 2026-09-05. Using the open deals and the engagements-by-deal table, list every open deal with no logged email, call, or meeting in the last 7 days. Compute recency from the engagements table's last_email/last_call/last_meeting fields, NOT from any date field on the deal (the deal's last_contacted_field is unreliable). For each stale deal give: deal alias, owner name, stage, amount, days since last contact. Group by owner, order by amount descending within each owner. Then state, per owner, how many stale deals they carry and the total stale amount. (Do not write outreach text.)
cost $0.2034 · 411s · in 43,649 / out 57,804 tok
✓ stale_recall_ge_80✓ grouped_by_owner✓ bryce_top_amount✓ cole_stale_amt✓ dana_stale_amt✓ stale_count

✓ Correct Answer (ground truth)

Plain-English answer: 65 open deals are stale; the largest is Deal-2D1F1B at $240,000, with Bryce Harmon carrying the most stale dollars at $626,243.
The answer in one line: jake_amt = 626,243; jenn_amt = 261,645; kevin_amt = 252,905.03; top_amount = 240,000; stale_count = 65; top_alias = Deal-2D1F1B
jake_amt626,243
jenn_amt261,645
kevin_amt252,905.03
stale_count65
top_amount240,000
top_aliasDeal-2D1F1B
graded checksstale_recall_ge_80 · grouped_by_owner · bryce_top_amount · cole_stale_amt · dana_stale_amt · stale_count

Model's actual answer

show the model's full answer
Method: snapshot 2026-09-05; "last 7 days" = 2026-08-29 through 2026-09-05. For each open deal, last contact = max(last_email, last_call, last_meeting) from engagements_by_deal_90d; stale if that date is before 2026-08-29 (days since > 7). All 156 rows of deals_open.csv treated as open.

STALE DEALS (65 total), grouped by owner, amount desc within owner:

BRYCE HARMON (13 stale)
- Deal-2D1F1B | DS1 | 240,000 | last 2026-06-16 | 81 days (14+31+31+5)
- Deal-66D1FC | DS1 | 99,000 | last 2026-08-20 | 16 days
- Deal-950043 | DS1 | 70,000 | last 2026-08-17 | 19 days
- Deal-B23205 | DS1 | 45,000 | last 2026-08-20 | 16 days
- Deal-7BBDFA | DS3 | 37,440 | last 2026-07-21 | 46 days (10+31+5)
- Deal-332637 | DS2 | 36,000 | last 2026-08-27 | 9 days
- Deal-1BEEBF | DS1 | 31,500 | last 2026-08-17 | 19 days
- Deal-C5658B | DS1 | 23,400 | last 2026-08-20 | 16 days
- Deal-40522D | DS3 | 21,000 | last 2026-08-17 | 19 days
- Deal-F0EBBB | DS3 | 11,400 | last 2026-08-12 | 24 days
- Deal-E25A09 | DS1 | 6,000 | last 2026-08-27 | 9 days
- Deal-C9C286 | DS2 | 5,502 | last 2026-08-27 | 9 days
- Deal-012CB1 | DS1 | 1 | last 2026-08-13 | 23 days
  Sum: 240,000+99,000+70,000+45,000+37,440+36,000+31,500+23,400+21,000+11,400+6,000+5,502+1 = 626,243

DANA MERCER (14 stale)
- Deal-44EA29 | DS2 | 60,000 | last 2026-08-26 | 10 days
- Deal-E51FB7 | DS2 | 43,875 | last 2026-08-24 | 12 days
- Deal-B42F46 | DS1 | 27,000 | last 2026-08-17 | 19 days
- Deal-BA3DDC | DS3 | 23,400 | last 2026-08-21 | 15 days
- Deal-9DDE86 | DS2 | 20,000 | last 2026-08-21 | 15 days
- Deal-215CCA | DS3 | 18,900 | last 2026-08-19 | 17 days
- Deal-5EED42 | DS3 | 16,250 | last 2026-08-25 | 11 days
- Deal-57887A | DS2 | 15,000 | last 2026-08-28 | 8 days
- Deal-B7EBD1 | DS5 | 9,000 | last 2026-08-20 | 16 days
- Deal-3974EB | DS4 | 9,000 | last 2026-08-28 | 8 days
- Deal-F40F04 | DS2 | 8,100 | last 2026-08-21 | 15 days
- Deal-87DDD1 | DS1 | 5,000 | last 2026-08-17 | 19 days
- Deal-F336B6 | DS3 | 4,200 | last 2026-08-21 | 15 days
- Deal-0660B4 | DS4 | 1,920 | last 2026-08-20 | 16 days
  Sum: 60,000+43,875+27,000+23,400+20,000+18,900+16,250+15,000+9,000+9,000+8,100+5,000+4,200+1,920 = 261,645

ALEX FRANKLIN (19 stale)
- Deal-CC08D1 | DS1 | 24,000 | last 2026-08-20 | 16 days
- Deal-E73427 | DS3 | 18,000 | last 2026-08-26 | 10 days
- Deal-885F45 | DS2 | 9,300 | last 2026-08-24 | 12 days
- Deal-C2FF3C | DS1 | 8,316 | last 2026-08-26 | 10 days
- Deal-3EED2C | DS2 | 7,200 | no engagement row | n/a
- Deal-0D2F7A | DS3 | 5,100 | last 2026-08-24 | 12 days
- Deal-6C60D4 | DS3 | 4,800 | last 2026-08-24 | 12 days
- Deal-13FEBD | DS2 | 4,680 | last 2026-08-24 | 12 days
- Deal-9D0060 | DS3 | 3,840 | last 2026-08-24 | 12 days
- Deal-690476 | DS2 | 3,600 | last 2026-08-18 | 18 days
- Deal-C6D97A | DS4 | 3,240 | last 2026-08-28 | 8 days
- Deal-EE195F | DS3 | 3,120 | last 2026-08-28 | 8 days
- Deal-278DEC | DS3 | 2,700 | last 2026-08-28 | 8 days
- Deal-635B8E | DS3 | 2,600 | last 2026-08-18 | 18 days
- Deal-6883F3 | DS1 | 2,400 | last 2026-08-20 | 16 days
- Deal-4A13AD | DS3 | 2,160 | last 2026-08-10 | 26 days
- Deal-F67D31 | DS2 | 1,800 | last 2026-08-28 | 8 days
- Deal-5FDCE4 | DS3 | 1,600 | last 2026-08-24 | 12 days
- Deal-BA571A | DS4 | 1,080 | last 2026-08-18 | 18 days
  Sum: 24,000+18,000+9,300+8,316+7,200+5,100+4,800+4,680+3,840+3,600+3,240+3,120+2,700+2,600+2,400+2,160+1,800+1,600+1,080 = 109,536

COLE INGRAM (18 stale)
- Deal-D04904 | DS2 | 58,529.25 | last 2026-08-25 | 11 days
- Deal-B25F40 | DS3 | 40,000 | last 2026-08-28 | 8 days
- Deal-813836 | DS2 | 32,175 | last 2026-08-25 | 11 days
- Deal-1BA595 | DS2 | 31,750 | last 2026-08-25 | 11 days
- Deal-CFE1E8 | DS3 | 18,000 | last 2026-08-25 | 11 days
- Deal-CD47A6 | DS2 | 12,168 | last 2026-08-25 | 11 days
- Deal-627646 | DS3 | 11,193 | last 2026-08-25 | 11 days
- Deal-FF809F | DS2 | 7,781.2 | last 2026-08-25 | 11 days
- Deal-AF932D | DS2 | 7,225.4 | last 2026-08-25 | 11 days
- Deal-A71728 | DS2 | 6,947.5 | last 2026-08-25 | 11 days
- Deal-8BC9F5 | DS2 | 5,616 | last 2026-08-26 | 10 days
- Deal-175395 | DS3 | 4,779.88 | last 2026-08-25 | 11 days
- Deal-481E24 | DS3 | 4,140 | last 2026-08-26 | 10 days
- Deal-C7F9BF | DS2 | 3,360 | last 2026-08-25 | 11 days
- Deal-2F3A66 | DS3 | 3,334.8 | last 2026-08-25 | 11 days
- Deal-342E96 | DS2 | 2,700 | last 2026-08-12 | 24 days
- Deal-E568D5 | DS3 | 1,875 | last 2026-08-25 | 11 days
- Deal-FD9F4E | DS5 | 1,330 | last 2026-08-26 | 10 days
  Sum = 252,905.03

ELENA SINCLAIR (1 stale)
- Deal-57FF13 | DS1 | 2,100 | no engagement row | n/a
  Sum: 2,100

FARID OSMAN (0 stale)
  Sum: 0

SUMMARY PER OWNER (stale count, stale amount):
- Bryce Harmon: 13 | 626,243
- Dana Mercer: 14 | 261,645
- Alex Franklin: 19 | 109,536
- Cole Ingram: 18 | 252,905.03
- Elena Sinclair: 1 | 2,100
- Farid Osman: 0 | 0
Total: 65 stale deals | 1,252,429.03 (626,243+261,645+109,536+252,905.03+2,100+0)

DATA NOTES (explicit, per instruction):
1. Missing data: two open deals have no row in engagements_by_deal_90d at all — Deal-3EED2C (Alex Franklin, 7,200) and Deal-57FF13 (Elena Sinclair, 2,100). With no email/call/meeting fields to evaluate, they carry no logged contact in the last 7 days and are counted stale; days-since-last-contact is not computable (n/a).
2. Future-dated engagements: 17 non-stale deals have a last_meeting dated after the snapshot. Ten of them are non-stale ONLY because of that future date (all their other last_* dates are >7 days old): Deal-944310 (Dana Mercer, 33 days), Deal-7599B8 (Dana Mercer, 18), Deal-3795AD (Bryce Harmon, 8), Deal-01E193 (Bryce Harmon, 8), Deal-C1FA6D (Bryce Harmon, 16), Deal-927338 (Bryce Harmon, 18), Deal-A414F6 (Bryce Harmon, 19), Deal-819506 (Alex Franklin, 8), Deal-8BA24E (Farid Osman, 8), Deal-8FDCD2 (Farid Osman, 15). Taking the engagements fields at face value (max of the three) they are not stale, as reported above. If you instead treat future-dated meetings as not logged, stale becomes 75 deals / 1,344,281.03: Bryce 18 / 692,964; Dana 16 / 279,495; Alex 20 / 113,936; Cole 18 / 252,905.03; Farid 2 / 2,881; Elena 1 / 2,100.
3. Archived owners (Gavin Porter, Hugo Lindqvist) have no deals in deals_open.csv. No deal's last contact fell exactly on the 2026-08-29 boundary, so the window-edge choice changes no result.
activity-mix-vs-outcomePASS 1.00
Plain-English question: Per rep, what did the last 30 days of emails, calls, and meetings look like next to the deals that advanced?
Prompt — what the model saw
The snapshot date is 2026-09-05. For each sales rep (owner), compute the last-30-day totals of emails, calls, and meetings from the engagements table, and the number of deals that entered DS2 in the last 30 days (t_ds2 within 30 days of the snapshot). Report per rep: the activity mix as percentages (emails/calls/meetings share of that rep's total activities), total activities per DS2 entry, and rank the reps by that efficiency ratio (lowest activities per DS2 entry = most efficient). Name the most efficient rep and the highest-volume rep, and note if they differ. Do not attribute cause.
cost $0.1616 · 280s · in 56,164 / out 41,341 tok
✓ alex_ds2_30d✓ alex_volume_high✓ efficient_named✓ mix_present✓ no_causal_claim

✓ Correct Answer (ground truth)

Plain-English answer: Alex Franklin is the standout signal with 18 deals entering DS2 in 30 days (next closest: Bryce Harmon with 4), alongside per-rep activity totals.
The answer in one line: alex_ds2_30d = 18; bryce_ds2_30d = 4; note = legacy alias justin_ds2=4 actually held Bryce Harmon count; Alex Franklin entered 18 deals into DS2 in the 30d window
alex_ds2_30d18
bryce_ds2_30d4
notelegacy alias justin_ds2=4 actually held Bryce Harmon count; Alex Franklin entered 18 deals into DS2 in the 30d window
graded checksalex_ds2_30d · alex_volume_high · efficient_named · mix_present · no_causal_claim

Model's actual answer

show the model's full answer
All figures computed from the three files as given. Window: t_ds2 between 2026-08-06 and 2026-09-05 inclusive (snapshot 2026-09-05 minus 30 days). Activity totals = sum of emails_30d + calls_30d + meetings_30d over each rep's open deals (all engagement rows join to a deal; no orphans).

Per-rep results

Alex Franklin (67 deals, 66 with engagement rows)
- E=307, C=36, M=41; total = 307+36+41 = 384
- Mix: 307/384 = 79.9% emails, 36/384 = 9.4% calls, 41/384 = 10.7% meetings
- DS2 entries (18): Deal-EE195F, Deal-D9A72E (08-06); Deal-7FA0C3, Deal-E531A6 (08-07); Deal-36C33F, Deal-D1E6C2 (08-11); Deal-317E6F (08-12); Deal-4F775F (08-17); Deal-F436DA (08-19); Deal-CA5E44 (08-24); Deal-5296C9, Deal-898FC5, Deal-E73427 (08-28); Deal-46988D (08-26); Deal-403845, Deal-92D97D (09-02); Deal-1FC049, Deal-3EED2C (09-03)
- Activities per DS2 entry: 384/18 = 21.33

Bryce Harmon (35 deals)
- E=162, C=0, M=43; total = 205
- Mix: 79.0% / 0.0% / 21.0%
- DS2 entries (4): Deal-1CCE5C (08-06), Deal-25F752 (08-10), Deal-CA7DC0 (08-12), Deal-D73B89 (09-03)
- 205/4 = 51.25

Cole Ingram (22 deals)
- E=96, C=14, M=1; total = 111
- Mix: 86.5% / 12.6% / 0.9%
- DS2 entries (2): Deal-1BA595 (08-12), Deal-42326B (08-26)
- 111/2 = 55.50

Farid Osman (7 deals)
- E=38, C=0, M=34; total = 72
- Mix: 52.8% / 0.0% / 47.2%
- DS2 entries (1): Deal-499BF6 (08-26)
- 72/1 = 72.00

Dana Mercer (24 deals)
- E=84, C=18, M=11; total = 113
- Mix: 74.3% / 15.9% / 9.7%
- DS2 entries (1): Deal-57887A (08-07)
- 113/1 = 113.00

Elena Sinclair (1 deal, Deal-57FF13)
- No row in the engagements file and no t_ds2 value. Activity and DS2 data are missing; mix and ratio cannot be computed.

Ranking by activities per DS2 entry (lowest = most efficient)

1. Alex Franklin — 21.33
2. Bryce Harmon — 51.25
3. Cole Ingram — 55.50
4. Farid Osman — 72.00
5. Dana Mercer — 113.00
(n/a: Elena Sinclair)

Most efficient rep: Alex Franklin (21.33 activities per DS2 entry).
Highest-volume rep: Alex Franklin (384 total activities, 79.9% emails / 9.4% calls / 10.7% meetings).
They do not differ — the same rep leads both.

Data caveats
- Deal-3EED2C (Alex Franklin, t_ds2 2026-09-03) has no engagement row, so it contributes 0 activities but counts as a DS2 entry. Excluding it would give 384/17 = 22.59; Alex stays #1 either way.
- Deal-57FF13 (Elena Sinclair) is absent from the engagements file entirely.
- Archived owners Gavin Porter and Hugo Lindqvist hold no deals in the snapshot and are excluded.
- No causal interpretation implied; ratios compare 30-day activity volume against same-window stage entries only.
ae-qtd-scorecardPASS 1.00
Plain-English question: Build Alex Franklin's QTD scorecard: bookings versus quota, pipeline, and conversion.
Prompt — what the model saw
Build a QTD scorecard for the named AE as of 2026-09-05: bookings vs quota with attainment percent, new vs expansion split, active pipeline by stage with amount, rolling 90-day DS2-to-won rate, win and loss counts with top loss reason, activity volume by type in the last 30 days. Exclude any closed-won deal dated before the quarter. Close with three coaching observations grounded in the numbers.
cost $0.1800 · 279s · in 48,259 / out 44,004 tok
✓ bookings✓ quota✓ attainment✓ new_expansion_split✓ top_loss_reason✓ coaching_obs

✓ Correct Answer (ground truth)

Plain-English answer: 75% attainment ($150,000 on a $200,000 quota), $113,500 new versus $36,500 expansion, pipeline by stage, and the rolling DS2-to-won rate.
The answer in one line: quota = 200000; bookings = 150000; new_bookings = 113500; expansion_bookings = 36500; open_deals_count = 125; attainment_pct = 75; owner = Alex Franklin (+12 more below)
quota200,000
attainment_pct75
bookings150,000
expansion_bookings36,500
expansion_count3
losses_qtd27
new_bookings113,500
new_count5
open_deals_count125
wins_qtd8
activity_30d_totalsemails_30d=807; calls_30d=112; meetings_30d=128; notes_30d=50
allowed_new18 items — e.g. Deal-A1C3E5, Deal-B7D2F4, Deal-C9E1A6, Deal-D4B8C2
ds2_to_wonwindow_start=2026-06-07; won_in_window=8; closed_in_window=35; rate=0.2286
ownerAlex Franklin
pipeline_by_stage{"DS1": {"count": 20, "amount": 284621}, "DS2": {"count": 28, "amount": 353760}, "DS3": {"count": 67, "amount": 552705}, "DS4": {"count": 5, "amount": 23574}, "DS5": {"count": 5, "amount": 45730}}
prior_quarter_excludeddeal_alias=Deal-B3E6F1; close_date=2026-06-20; amount=24000
quarter2026-Q3
snapshot_date2026-09-05
top_loss_reasonreason=Lost- Timing (1 year or more); count=13
graded checksbookings · quota · attainment · new_expansion_split · top_loss_reason · coaching_obs

Model's actual answer

show the model's full answer
# QTD Scorecard — Alex Franklin, as of 2026-09-05 (2026-Q3)

## 1. Bookings vs quota
| Item | Value |
|---|---|
| QTD bookings (closed-won 2026-07-01 → 2026-09-05) | $150,000 |
| Q3 quota | $200,000 |
| Attainment | 150,000 / 200,000 = **75.0%** |

Arithmetic: 40,000 + 20,000 + 35,000 + 21,000 + 12,000 + 11,000 + 6,500 + 4,500 = 150,000
Excluded per rule (closed-won before quarter): **Deal-B3E6F1**, $24,000, closed 2026-06-20 — excluded from bookings.

QTD wins by deal: Deal-A1C3E5 $40,000 (new, 07-15); Deal-F2C7D8 $20,000 (expansion, 07-24); Deal-B7D2F4 $35,000 (new, 07-31); Deal-C9E1A6 $21,000 (new, 08-12); Deal-A8B4D6 $12,000 (expansion, 08-19); Deal-D4B8C2 $11,000 (new, 08-21); Deal-E6F3A9 $6,500 (new, 09-02); Deal-C5D9E2 $4,500 (expansion, 09-03).

## 2. New vs expansion split
| Type | Count | Amount | Share |
|---|---|---|---|
| New | 5 | $113,500 | 75.7% |
| Expansion | 3 | $36,500 | 24.3% |
| Total | 8 | $150,000 | 100% |

Arithmetic: new = 40,000+35,000+21,000+11,000+6,500 = 113,500; expansion = 20,000+12,000+4,500 = 36,500; 113,500+36,500 = 150,000 ✓

## 3. Active pipeline by stage (status=open, 125 deals)
| Stage | Deals | Amount |
|---|---|---|
| DS1 | 20 | $284,621 |
| DS2 | 28 | $353,760 |
| DS3 | 67 | $552,705 |
| DS4 | 5 | $23,574 |
| DS5 | 5 | $45,730 |
| **Total** | **125** | **$1,260,390** |

Arithmetic: 284,621+353,760+552,705+23,574+45,730 = 1,260,390 ✓
Subsets: 22 open deals ($109,363) carry a Q3 close date (2026-07-01→09-30); only 10 deals ($69,304) sit in DS4–DS5. Data flag: Deal-7A2454 (DS3, $1,275) is still open with close_date 2026-09-04, before the as-of date — stale date, kept in pipeline since status=open.

## 4. Rolling 90-day DS2-to-won rate
Cohort = deals with entered_ds2 in 2026-06-07 → 2026-09-05 (90 days back from as-of): **111 deals** (8 won, 27 lost, 76 still open).
- Won / cohort: 8 / 111 = **7.2%**
- Won / decided (won+lost): 8 / 35 = **22.9%**
Note: Deal-B3E6F1 is the only closed-won outside this cohort (entered_ds2 2026-05-10); all 8 QTD wins are in it.

## 5. Win / loss counts and top loss reason (QTD, closed 07-01 → 09-05)
- Wins: 8 ($150,000) | Losses: 27 ($329,272)
- Top loss reason: **Lost- Timing (1 year or more)** — 13 deals, $184,681 (48% of losses by count, 56% of lost dollars)

| Loss reason | Deals | Amount |
|---|---|---|
| Lost- Timing (1 year or more) | 13 | $184,681 |
| MIA | 5 | $45,831 |
| Competitor | 5 | $49,020 |
| Lost DM | 2 | $17,940 |
| Feature Request | 1 | $21,000 |
| Lost- Does not fit ICP (write in notes) | 1 | $10,800 |

## 6. Activity volume, last 30 days (per-deal 30d counts, as of snapshot)
| Type | Total |
|---|---|
| Emails | 807 |
| Calls | 112 |
| Meetings | 128 |
| Notes | 50 |
| **Total touches** | **1,097** |

Split by deal status: open $599 / 54 / 90 / 1 (em/co/me/no); lost 109 / 25 / 13 / 25; won 99 / 33 / 25 / 24. Data caveat: the engagement file carries no dates, only 30d counters per deal; treated as the window ending at the 2026-09-05 snapshot.

## 7. Three coaching observations
1. **In-quarter coverage is thin relative to win rate.** Gap to quota is $50,000 (200,000−150,000), but only $109,363 of open pipeline carries a Q3 close date — 2.19x coverage, and only 10 deals ($69,304) are in DS4/DS5. At the 22.9% decided win rate (8/35), that Q3-dated pipeline yields roughly $25,000 expected — short of the $50,000 gap. The 125-deal / $1.26M book is real coverage only for later quarters (115 of 125 deals sit in DS1–DS3).
2. **Timing is the dominant leak and it is expensive.** "Lost- Timing (1 year or more)" is 13 of 27 losses (48%) and $184,681 of $329,272 lost dollars (56%) — more than MIA and Competitor combined ($94,851). Deals with no committed buying timeline are being carried into DS2+ rather than qualified or parked early (e.g., Deal-B3ABED $40,001, Deal-E6E80A $24,000, Deal-79B7A1 $25,000 all lost on timing after entering DS2 in July).
3. **Activity is email-heavy and that correlates with losing.** The 30d mix is 807 emails vs 112 calls and 128 meetings (email = 74% of all touches); 103 of 125 open deals had zero calls and 82 of 125 had zero meetings in the window. By contrast, the 8 QTD wins averaged 3.9 calls and 2.9 meetings per deal vs 0.9 calls and 0.5 meetings on the 27 QTD losses. The coaching lever is multi-channel contact on open pipeline — emails alone are not holding deals (MIA is still 5 losses / $45,831), and Deal-3EED2C shows zero recorded activity of any kind while open.

Source files: ae_deals.csv (161 rows), ae_engagements.csv (161 rows), quota.csv; all figures computed from these only.

⚠️ File-mutation verifier: 2 file edit(s) FAILED this turn despite any wording above that may suggest otherwise. Run `git status` or `read_file` to confirm what actually landed.
  • `/Users/amaniphipps/.hermes/cache/scratch/ae_deals.csv` — [write_file] Refusing to overwrite `/Users/amaniphipps/.hermes/cache/scratch/ae_deals.csv`: `/Users/amaniphipps/.hermes/cache/scratch/ae_deals.csv` exists but this task has not seen its full curre…
  • `/Users/amaniphipps/.hermes/cache/scratch/ae_engagements.csv` — [write_file] Refusing to overwrite `/Users/amaniphipps/.hermes/cache/scratch/ae_engagements.csv`: `/Users/amaniphipps/.hermes/cache/scratch/ae_engagements.csv` exists but this task has not seen it…
multithreading-gapPASS 1.00
Plain-English question: Which open deals are single-threaded or under-threaded?
Prompt — what the model saw
Find every open deal that is single-threaded (fewer than 2 active contacts) or under-threaded (fewer than 3, or all contacts in one persona). Active means engaged in the last 60 days and not marked former. Personas: economic buyer, champion, HR admin, IT security, finance. For each flagged deal: amount, stage, active contact count, personas present, personas missing, the most valuable persona to add given the stage, and any on-file unengaged contact at that company who fits; otherwise say none on file.
cost $0.0351 · 87s · in 3,026 / out 10,711 tok
✓ flagged_recall✓ persona_language✓ single_thread_term✓ on_file_fit

✓ Correct Answer (ground truth)

Plain-English answer: Deals with fewer than 2 active contacts (single-threaded) or fewer than 3, or all in one persona (under-threaded), each flagged with the missing persona.
The answer in one line: active_window_start = 2026-07-07
active_window_start2026-07-07
allowed_new59 items — e.g. C-077A0E, C-0D15DF, C-1BDFE0, C-2AE3AA
best_persona_ruleDS2/DS3: economic buyer > IT security > finance > HR admin > champion; DS4/DS5: economic buyer > finance > IT security > HR admin > champion; first persona not present among active contacts
flag_ruleflag if active_contacts < 2 (single-threaded) or active_contacts < 3 or all active contacts share one persona (under-threaded); active = engaged within 60 days and not former
flaggeddeal_alias=Deal-36C33F; amount=15000; stage=DS2; active_count=1; reason=single-threaded; best_persona_to_add=economic buyer | deal_alias=Deal-50D386; amount=36000; stage=DS2; active_count=2; reason=under-threaded; best_persona_to_add=economic buyer | deal_alias=Deal-5408B0; amount=14850; stage=DS4; active_count=2; reason=under-threaded; best_persona_to_add=economic buyer | deal_alias=Deal-5BFE3B; amount=18000; stage=DS2; active_count=2; reason=under-threaded; best_persona_to_add=economic buyer | deal_alias=Deal-885F45; amount=9300; stage=DS2; active_count=2; reason=under-threaded; best_persona_to_add=IT security | deal_alias=Deal-92D97D; amount=60000; stage=DS2; active_count=1; reason=single-threaded; best_persona_to_add=economic buyer … (+5 more)
healthy_not_flaggedDeal-4B0BEB, Deal-84DBA6, Deal-D348E1
persona_universeeconomic buyer, champion, HR admin, IT security, finance
graded checksflagged_recall · persona_language · single_thread_term · on_file_fit

Model's actual answer

show the model's full answer
ASSUMPTIONS AND DERIVED VALUES

- As-of date: 2026-10-19 (session date). 60-day cutoff = 2026-10-19 minus 60 days = 2026-08-20 (Oct 19 to Sep 19 = 30 days; Sep 19 to Aug 20 = 30 days).
- Active = is_former = false AND last_engaged_date >= 2026-08-20 (inclusive). Contacts at exactly 60 days (2026-08-20) counted active; sensitivity noted at bottom.
- Data gap, applies to every deal: neither file contains deal amount or deal stage. Amount and stage cannot be provided for any deal. Because stage is missing, the "most valuable persona given the stage" cannot be stage-conditioned; I give a stage-independent default priority instead: economic buyer if missing, else champion if missing, else any missing persona.
- 14 deals on file; 11 flagged, 3 clean.

FLAGGED DEALS (11)

1. Deal-EC3025 (C-FDD0C7) - single-threaded
   Amount: not in data. Stage: not in data.
   Active: 1 of 2 on file. CT-047C54, champion, 2026-09-02 (47d). Excluded: CT-F2C1AE, economic buyer, 2026-08-15 (65d) AND is_former=true.
   Personas present: champion. Missing: economic buyer, HR admin, IT security, finance.
   Best persona to add: economic buyer (stage-independent default; stage not on file).
   On-file unengaged match: CT-6827DB, Chief People Officer, economic buyer.

2. Deal-92D97D (C-E23238) - single-threaded
   Amount: not in data. Stage: not in data.
   Active: 1 of 2. CT-01F5B4, HR admin, 2026-08-28 (52d). Excluded: CT-A902AE, champion, 2026-06-01 (140d).
   Personas present: HR admin. Missing: economic buyer, champion, IT security, finance.
   Best persona to add: economic buyer (default; no stage).
   On-file unengaged match: none on file.

3. Deal-50D386 (C-EB10E4) - under-threaded (2 < 3)
   Amount: not in data. Stage: not in data.
   Active: 2 of 2. CT-AA41B2, champion, 2026-09-01 (48d); CT-B9C35B, HR admin, 2026-08-25 (55d).
   Personas present: champion, HR admin. Missing: economic buyer, IT security, finance.
   Best persona to add: economic buyer (default; no stage).
   On-file unengaged match: CT-A1C4B3, Chief People Officer, economic buyer.

4. Deal-D0D6B5 (C-32918E) - single-threaded
   Amount: not in data. Stage: not in data.
   Active: 1 of 3. CT-87CED4, champion, 2026-09-02 (47d). Excluded: CT-DE6D7C, champion, 2026-08-19 (61d); CT-FD70B2, champion, 2026-08-07 (73d).
   Personas present: champion. Missing: economic buyer, HR admin, IT security, finance.
   Best persona to add: economic buyer (default; no stage).
   On-file unengaged match: CT-1FA4DB, Chief People Officer, economic buyer.

5. Deal-5BFE3B (C-535D36) - single-threaded
   Amount: not in data. Stage: not in data.
   Active: 1 of 2. CT-57123B, champion, 2026-08-31 (49d). Excluded: CT-5CE757, champion, 2026-08-12 (68d).
   Personas present: champion. Missing: economic buyer, HR admin, IT security, finance.
   Best persona to add: economic buyer (default; no stage).
   On-file unengaged match: none on file.

6. Deal-36C33F (C-077A0E) - zero active contacts (fewer than 2)
   Amount: not in data. Stage: not in data.
   Active: 0 of 3. Excluded: CT-4FE556, IT security, 2026-08-15 (65d, outside window); CT-405B45, champion, former; CT-86B22F, economic buyer, former.
   Personas present: none. Missing: economic buyer, champion, HR admin, IT security, finance (all five).
   Best persona to add: economic buyer (default; no stage).
   On-file unengaged match: CT-1DB73E, Chief People Officer, economic buyer.

7. Deal-885F45 (C-5E8EFB) - single-threaded
   Amount: not in data. Stage: not in data.
   Active: 1 of 2. CT-51C81E, economic buyer, 2026-08-26 (54d). Excluded: CT-D9A0E8, champion, 2026-08-11 (69d).
   Personas present: economic buyer. Missing: champion, HR admin, IT security, finance.
   Best persona to add: champion (economic buyer already present; default order; no stage).
   On-file unengaged match: no champion on file. One unengaged contact fills a different missing persona: CT-B3F25D, IT Security Lead, IT security.

8. Deal-FCBE5B (C-737030) - single-threaded
   Amount: not in data. Stage: not in data.
   Active: 1 of 1. CT-4A5317, champion, 2026-08-29 (51d).
   Personas present: champion. Missing: economic buyer, HR admin, IT security, finance.
   Best persona to add: economic buyer (default; no stage).
   On-file unengaged match: none on file.

9. Deal-5408B0 (C-2AE3AA) - single-threaded
   Amount: not in data. Stage: not in data.
   Active: 1 of 2. CT-D33AE4, champion, 2026-09-01 (48d). Excluded: CT-8742FD, HR admin, 2026-08-18 (62d).
   Personas present: champion. Missing: economic buyer, HR admin, IT security, finance.
   Best persona to add: economic buyer (default; no stage).
   On-file unengaged match: CT-07FA76, Chief People Officer, economic buyer.

10. Deal-C6D97A (C-5A8FC2) - under-threaded (2 < 3) AND all contacts in one persona
    Amount: not in data. Stage: not in data.
    Active: 2 of 3, both champion. CT-223DDC, 2026-08-31 (49d); CT-B03555, 2026-08-20 (60d, boundary). Excluded: CT-4E8A2B, champion, 2026-08-05 (75d).
    Personas present: champion only. Missing: economic buyer, HR admin, IT security, finance.
    Best persona to add: economic buyer (default; no stage).
    On-file unengaged match: none on file.

11. Deal-F9A08A (C-0D15DF) - single-threaded
    Amount: not in data. Stage: not in data.
    Active: 1 of 2. CT-931B10, champion, 2026-09-03 (46d). Excluded: CT-913581, economic buyer, 2026-06-20 (121d).
    Personas present: champion. Missing: economic buyer, HR admin, IT security, finance.
    Best persona to add: economic buyer (default; no stage).
    On-file unengaged match: CT-697541, Chief People Officer, economic buyer.

NOT FLAGGED (3)

- Deal-4B0BEB (C-1BDFE0): 3 active, 3 personas - CT-A96531 champion 2026-09-01 (48d), CT-DE5BDC economic buyer 2026-08-28 (52d), CT-1543C8 HR admin 2026-08-21 (59d). Excluded: CT-3E135F finance 2026-08-09 (71d). Passes both thresholds.
- Deal-84DBA6 (C-BE9BE0): 3 active, 3 personas - CT-BAA9D3 champion 2026-09-02 (47d), CT-60834D economic buyer 2026-08-30 (50d), CT-3F23A4 IT security 2026-08-20 (60d, boundary). Passes only because the boundary contact counts.
- Deal-D348E1 (C-804F29): 4 active, 4 personas - CT-4EA0A4 champion 2026-09-03 (46d), CT-2164AD economic buyer 2026-09-02 (47d), CT-EC404C IT security 2026-08-30 (50d), CT-08E5FA finance 2026-08-27 (53d). Excluded: CT-BEDF5E HR admin 2026-08-19 (61d).

BOUNDARY SENSITIVITY

Two contacts sit at exactly 60 days (2026-08-20): CT-3F23A4 (Deal-84DBA6, IT security) and CT-B03555 (Deal-C6D97A, champion). If the window is read as strictly under 60 days, Deal-84DBA6 drops to 2 active (champion, economic buyer) and becomes under-threaded with IT security missing (its only on-file unengaged contact, CT-2A0169 Controller/finance, fills finance, also missing); Deal-C6D97A drops to 1 active and is flagged either way. Everything else is unaffected.

STILL MISSING FOR A FULL ANSWER

Deal amount and stage for all 14 deals, and any stage-conditional persona guidance, are not in the provided files. Add deal amount and stage and I can finish the per-deal economics and the stage-weighted persona recommendation.
talk-track-from-transcriptsPASS 1.00
Plain-English question: From this rep's last 10 calls: how do they open, handle objections, and how often do they lock a next step?
Prompt — what the model saw
From this rep's last 10 calls: what they lead with in the first five minutes, how they handle the three most common objections, how often a concrete next step is agreed (rate), and every competitor a prospect raised. Quote at most one line per finding and only lines that appear in the transcripts. End with two coaching notes.
cost $0.0230 · 58s · in 3,872 / out 6,510 tok
✓ objection_count✓ next_step_rate✓ competitors✓ coaching✓ quote_used

✓ Correct Answer (ground truth)

Plain-English answer: 8 of 10 calls open with the same customer-win story; objection handling, next-step agreement rate, and every competitor raised, with quotes.
The answer in one line: n_transcripts = 10; objection_count = 3; rep = Alex Franklin (+3 more below)
n_transcripts10
objection_count3
allowed_new13 items — e.g. TT-001, TT-002, TT-003, TT-004
competitor_quotesAwardco=We're also in late talks with Awardco — their rewards catalog looks bigger than yours.; Kudos=How are you different from Kudos? Our CEO used them at her last company.
competitors_prospect_raisedAwardco, Kudos
next_stepagreed=7; total=10; rate=0.7
objectionsobjection=budget; count=4; quote=Honestly, budget is locked until next fiscal year — I can't add a new line item right now.; transcript_id=TT-001 | objection=timing; count=3; quote=This is good, but can we revisit it next quarter? Open enrollment starts in October and we're underwater.; transcript_id=TT-002 | objection=in-house; count=3; quote=We already do recognition with a spreadsheet and quarterly gift cards — why would we change?; transcript_id=TT-004
opener_themetheme=opens with a customer recognition win story (400-person retailer, 18% turnover reduction); count=8; total=10; quote=Quick one to kick off: a 400-person retailer we work with cut regretted turnover 18% in two quarters after automating milestone awards, and I wanted to walk you through exactly how they did it.; transcript_id=TT-001
repAlex Franklin
rep_raised_competitors_excludedWorkhuman
graded checksobjection_count · next_step_rate · competitors · coaching · quote_used

Model's actual answer

show the model's full answer
FIRST FIVE MINUTES - WHAT THE REP (ALEX FRANKLIN) LEADS WITH

- 8 of 10 calls (TT-001/Deal-D348E1, TT-002/Deal-5408B0, TT-003/Deal-547B2B, TT-005/Deal-C61CF7, TT-006/Deal-60C2C2, TT-007/Deal-EDC141, TT-008/Deal-D9A12F, TT-010/Deal-84DBA6) open with the identical retailer proof point, word for word: "Quick one to kick off: a 400-person retailer we work with cut regretted turnover 18% in two quarters after automating milestone awards, and I wanted to walk you through exactly how they did it."
- TT-004/Deal-403845 opens with an agenda instead: "I put together a short agenda — security review first, then pricing."
- TT-009/Deal-1E2498 opens with pricing per the prior request: "You asked for straight pricing last time, so let's start there."
- TT-005/Deal-C61CF7 adds a rep-initiated competitor comparison at minute 2: "And unlike Workhuman, our pricing includes the full rewards catalog with no extra margin."
- Arithmetic: 8 story + 1 agenda + 1 pricing = 10 calls.

TOP THREE OBJECTIONS AND HOW HE HANDLES THEM

1. Budget locked (4 of 10: TT-001, TT-003, TT-006, TT-010). He reframes funding to turnover savings with the same $210k figure every time: "Totally fair. Most teams fund this out of turnover savings — that retailer saved about $210k in avoided backfills, which is how their finance team signed off."
2. Defer to next quarter / open enrollment workload (3 of 10: TT-002, TT-005, TT-008). He counters with a scoped pilot to build internal data: "Makes sense. What if we scope a 90-day pilot with one department so you have internal data before next quarter's planning?"
3. Status quo, spreadsheet and gift cards (3 of 10: TT-004, TT-007, TT-009). He answers with automation and analytics: "Spreadsheets work until they scale — the difference is automation: milestones fire without HR lifting a finger, and you get analytics on who is being recognized."
- Arithmetic: budget 4 + timing 3 + status quo 3 = 10 of 15 total prospect objections; remaining 5 are committee gates (TT-004, TT-010), "no urgency" (TT-007), and the two competitor mentions (TT-003, TT-007).

CONCRETE NEXT STEP AGREED - RATE

- He asks for the next step at minute 14 in 7 of 10 calls (TT-001, TT-002, TT-003, TT-005, TT-006, TT-008, TT-009): "Should we lock the next step — a working session with your team this week?"
- The prospect accepts with a specific date, time, and attendee in all 7 asked: "Yes, Thursday at 2pm works — send the invite and I'll bring our HRIS manager."
- Rate: 7/10 = 70% of all calls end with a concrete next step; 7/7 = 100% of asks convert.
- The 3 misses (TT-004, TT-007, TT-010) never reach the ask; each ends after a committee or "need to think" pushback with a passive close, e.g. TT-004: "Understood — I'll leave it with you."

COMPETITORS RAISED BY PROSPECTS

- Awardco, TT-003/Deal-547B2B, minute 4: "We're also in late talks with Awardco — their rewards catalog looks bigger than yours."
- Kudos, TT-007/Deal-EDC141, minute 4: "How are you different from Kudos? Our CEO used them at her last company."
- Workhuman appears in TT-005/Deal-C61CF7 but was raised by the rep at minute 2, not by the prospect, so it does not count as prospect-raised. No other prospect-raised competitors appear in the 10 transcripts.

COACHING NOTES

1. Fix the close on committee and "no urgency" pushback. The 70% next-step rate is entirely a function of whether he asks: 7/7 asks convert, and the only 3 losses (TT-004, TT-007, TT-010) come when he drops the ask and ends passively ("Fair enough."). Build the same scripted counter he has for budget and timing objections for "the committee decides" and "I need to think about it," such as a decision-process check or a booked follow-up date, so the next-step ask lands in 10 of 10.
2. Diversify proof points and work the competitive moments. The same 18%/$210k retailer story is both the opener in 8 of 10 calls and the budget answer in 4 of 4, and competitive replies stay as one-line comparisons, e.g. TT-003: "Awardco's catalog is broad; where we win is automation and the analytics your HRIS team asked about." After a prospect names a competitor, especially "late talks with Awardco," probe the evaluation criteria and set a competitive next step instead of defaulting only to the generic working-session ask, and rotate the proof point so it doesn't read as scripted.
reporting-analytics 5 tests
q3-weighted-forecastPASS 1.00
Plain-English question: Build the Q3 2026 weighted forecast from the open pipeline, and show the math deal by deal.
Prompt — what the model saw
Below are the open deals in the default sales pipeline with a close date between 2026-07-01 and 2026-10-15, pulled from the CRM warehouse on 2026-09-05. Stages: DS1 (discovery) through DS5 (contract). Build the Q3 2026 forecast (quarter = 2026-07-01 to 2026-09-30). Weighting rule: 100 percent of COMMIT plus 35 percent of BEST_CASE. PIPELINE deals count zero. Only deals with a close date inside the quarter count. Report: COMMIT total, BEST_CASE total, weighted forecast, count of deals in each category inside the quarter, the deals you excluded for being outside the quarter (count and total amount), and the top 5 BEST_CASE deals by amount inside the quarter. Show arithmetic. Then, in a final '## Data quality' section of no more than four sentences, name the data quality problems in this extract that would make the forecast unreliable if it were run unattended.
cost $0.0613 · 107s · in 29,402 / out 14,083 tok
✓ commit_total_44729✓ best_case_total_203565✓ weighted_115976.75✓ commit_count_7✓ best_case_count_24✓ excluded_count_32✓ excluded_amount_227575✓ top5_aliases✓ flags_blank_owner✓ flags_missing_why_buys

✓ Correct Answer (ground truth)

Plain-English answer: A stage-weighted Q3 forecast number computed from the open pipeline, with per-stage math that reconciles to the deal data.
The answer in one line: excluded_amount = 227575; best_case_total = 203565; weighted_forecast = 115,976.75; commit_total = 44729; excluded_out_of_quarter = 32; best_case_count = 24; flags = blank owner column + empty why_buys fields must be surfaced (+2 more below)
excluded_amount227,575
best_case_count24
best_case_total203,565
commit_count7
commit_total44,729
excluded_out_of_quarter32
weighted_forecast115,976.75
flagsblank owner column + empty why_buys fields must be surfaced
top5_best_caseDeal-2D7423, Deal-25F752, Deal-E53952, Deal-5EED42, Deal-FA32A0
graded checkscommit_total_44729 · best_case_total_203565 · weighted_115976.75 · commit_count_7 · best_case_count_24 · excluded_count_32 · excluded_amount_227575 · top5_aliases · flags_blank_owner · flags_missing_why_buys

Model's actual answer

show the model's full answer
# Q3 2026 Forecast (2026-07-01 to 2026-09-30)

## In-quarter totals
- COMMIT: 7 deals, $44,729
  - 11,200 (Deal-547B2B) + 9,000 (Deal-B7EBD1) + 9,000 (Deal-403845) + 6,360 (Deal-A2B47C) + 5,400 (Deal-2465CE) + 2,520 (Deal-A5E80A) + 1,249 (Deal-499BF6) = 44,729
- BEST_CASE: 24 deals, $203,565
- PIPELINE: 23 deals, $201,637.40 — weighted at zero per rule
- Total in-quarter deals: 7 + 24 + 23 = 54 of 86 extract rows

## Weighted forecast
- Weighted = COMMIT + 0.35 × BEST_CASE
- 0.35 × 203,565 = 71,247.75
- Weighted = 44,729 + 71,247.75 = **$115,976.75**

## Excluded — close date outside the quarter (all after 2026-09-30)
- Count: 32 deals, total $227,575
- By category: COMMIT 1 ($13,770 — Deal-D348E1, 2026-10-15); BEST_CASE 9 ($28,240); PIPELINE 22 ($185,565)
- Check: 13,770 + 28,240 + 185,565 = 227,575; in-quarter 44,729 + 203,565 + 201,637.40 + 227,575 = 677,506.40 = extract total

## Top 5 BEST_CASE deals in quarter
1. Deal-2D7423 — $38,935 (DS3, close 2026-09-30)
2. Deal-25F752 — $24,000 (DS4, close 2026-09-25)
3. Deal-E53952 — $19,656 (DS4, close 2026-09-30)
4. Deal-5EED42 — $16,250 (DS3, close 2026-09-30)
5. Deal-FA32A0 — $11,116 (DS3, close 2026-09-25)
- Top-5 sum: 38,935 + 24,000 + 19,656 + 16,250 + 11,116 = 109,957 (54.0% of in-quarter BEST_CASE)

## Data quality
Owner is blank on 85 of 86 deals (only Deal-C9C286 has one), so forecast accountability and rep-level sanity checks are impossible from this extract. Stage/category contradictions exist: Deal-A5E80A is COMMIT at DS1 and Deal-499BF6 is COMMIT at DS2, which inflates the 100%-weighted bucket if stage progression is trustworthy. Why-buys is empty (0 chars) on 71 of 86 deals, and four PIPELINE deals have close dates already past the 2026-09-05 extract date (e.g., Deal-333EBB 2026-08-28), indicating stale or unmaintained records; additionally, 7 BEST_CASE deals ($96,257, ~47% of the BEST_CASE bucket) all close on exactly 2026-09-30, a quarter-end clustering pattern that risks over-weighting deals that will not actually close in quarter.
aha-moment-2x20.83
Plain-English question: Do companies that hit 5+ givers and one redemption in month one actually retain better at 24 months?
Prompt — what the model saw
Test the activation hypothesis on the company cohort below: companies with 5+ unique givers (m1_users >= 5) AND 1+ successful redemption (m1_redemptions >= 1) in their first calendar month retain better at 24 months than those with one signal or neither. Every company in this extract is already 25+ months old, so current_status = 'active' means retained at 24 months. Report the full 2x2 (cohort size and 24-month retention rate for: both signals, givers-only, redemption-only, neither), state how many companies were excluded from the denominator and why (if none, say so), name the single signal with the largest retention lift, and state what this does and does not prove.
cost $0.0668 · 117s · in 25,905 / out 15,272 tok
✓ both_66✓ givers_47✓ redemp_31✓ neither_40✗ excluded_0✓ largest_lift_givers

✓ Correct Answer (ground truth)

Plain-English answer: The four cohort cells are graded against computed 24-month retention; the numbers come from the scorer's cohort table, not opinion.
The answer in one line: both_5plus_givers_and_redemption = 66.0% (31 of 47)
both_5plus_givers_and_redemption66.0% (31 of 47)
exclusionsnone — claim bounded as correlation, not causation
givers_only46.9% (23 of 49)
largest_single_signal_liftunique givers
neither40.0% (38 of 95)
redemption_only31.0% (9 of 29)
graded checksboth_66 · givers_47 · redemp_31 · neither_40 · excluded_0 · largest_lift_givers

Model's actual answer

show the model's full answer
## Activation hypothesis test — 2023 PLG cohort (n = 220)

Definitions applied exactly as given: givers signal = m1_users ≥ 5; redemption signal = m1_redemptions ≥ 1; retained = current_status = 'active' (all companies 25+ months old; signup months 2023-01 through 2023-07).

### Full 2x2

| Cell | Cohort size | Retained (active) | 24-mo retention |
|---|---|---|---|
| Both signals | 47 | 31 | 31/47 = 66.0% |
| Givers-only (≥5 users, 0 redemptions) | 49 | 23 | 23/49 = 46.9% |
| Redemption-only (≥1 redemption, <5 users) | 29 | 9 | 9/29 = 31.0% |
| Neither | 95 | 38 | 38/95 = 40.0% |

Check: 47+49+29+95 = 220; 31+23+9+38 = 101 = total 'active' in file. Consistent.

Key contrasts vs. the neither baseline:
- Both vs neither: 66.0% − 40.0% = +26.0 pp
- Givers-only vs neither: 46.9% − 40.0% = +6.9 pp
- Redemption-only vs neither: 31.0% − 40.0% = −9.0 pp (negative, not positive)

### Exclusions from denominator
Zero. All 220 rows have non-missing m1_users, m1_redemptions, and current_status, and every signup month (2023-01 … 2023-07) is ≥25 months before the 2026-10-19 extract date, so nothing was excluded. Note: 3 companies with current_status = 'non_renewing' (C-0B2078FB in neither; C-0A96134F and C-0BEAF685 in redemption-only) are counted as NOT retained, since only 'active' is defined as retained — they remain in the denominators.

### Largest single-signal lift: the givers signal (m1_users ≥ 5)

Marginal (collapsed) comparison:
- Givers signal present (both + givers-only): 54/96 = 56.25% retained
- Givers signal absent (redemption-only + neither): 47/124 = 37.90% retained
- Lift = +18.3 pp
- Redemption signal present (both + redemption-only): 40/76 = 52.63%
- Redemption signal absent (givers-only + neither): 61/144 = 42.36%
- Lift = +10.3 pp

The same ordering holds in the isolated-cell view (givers-only +6.9 pp vs neither; redemption-only −9.0 pp vs neither). So breadth of giving is the stronger signal; redemption alone adds less and, without the giver threshold, associates with worse-than-baseline retention.

### What this does and does not prove

Does support: within this cohort, companies with both first-month signals retained materially better at 24+ months (66.0%) than companies with neither (40.0%), and better than either signal alone — directionally consistent with the activation hypothesis. Among single signals, the givers signal carries the largest lift (+18.3 pp marginal).

Does not prove:
- Causality. This is observational: companies that reached 5+ givers and a redemption in month one may have simply been larger, more committed, or better-resourced at signup. The extract has no company-size, headcount, or plan-eligibility field to control for this, so the lift may be selection, not activation effect.
- That driving redemptions causes retention. Redemption-only companies retained WORSE than the neither baseline (31.0% vs 40.0%), so a redemption without breadth of giving is not, in this data, a positive standalone signal — the both-vs-either gap is driven mostly by the givers threshold.
- Anything about cells beyond these thresholds (e.g., 5 vs 30 givers), about later-month behavior (m3 columns ignored by design), or about why non_renewing companies (3, all with nonzero current_arr_usd) are excluded from the retained count.
- A precise 24-month measurement: 'active' is a current status at 25+ months (27–45 months depending on signup month), not a snapshot taken exactly at month 24; and the redemption-only cell (n=29) is small enough that its 31.0% is imprecise.

No facts beyond the extract were used.
arr-rollforward-reconciliationPASS 1.00
Plain-English question: Does CRM ARR match billing ARR — and if not, exactly where is the variance?
Prompt — what the model saw
Reconcile ARR between CRM company records and billing subscriptions as of 2026-09-05. Billing ARR is MRR x 12. Report both totals, the variance, and a decomposition into buckets (status mismatch, rounding, missing records, other) that sums exactly to the variance. List each mismatched account with a suggested owner. Business rule: any subscription with a term other than 12 months must have cf_agreement_end_date populated; list violations.
cost $0.0571 · 137s · in 7,554 / out 16,739 tok
✓ billing_total✓ crm_total✓ variance✓ buckets_sum✓ violations_listed

✓ Correct Answer (ground truth)

Plain-English answer: Billing $604,739.28 vs CRM $603,581.76: a $1,157.52 variance decomposed into status-mismatch, rounding, and missing-record buckets.
The answer in one line: n_companies = 40; n_mismatched_accounts = 7; as_of = 2026-09-05 (+4 more below)
n_companies40
n_mismatched_accounts7
as_of2026-09-05
billing_arr_total604739.28
bucket_sum_equals_variance1
buckets{"status_mismatch": {"amount": "-13158.48", "accounts": [{"company_alias": "C-0C8323BF", "billing_arr": "0.00", "hubspot_arr": "4905.24", "diff": "-4905.24", "suggested_owner": "Dana Mercer"}, {"compa
crm_arr_total603581.76
ruleterm_months != 12 requires cf_agreement_end_date
variance1157.52
variance_definitionbilling_arr_total - crm_arr_total (billing ARR = sum of MRR x 12 over status=active subscriptions)
violationssubscription_id=SUB-0002; company_alias=C-1794A52C; term_months=24 | subscription_id=SUB-0019; company_alias=C-22170CA1; term_months=36
graded checksbilling_total · crm_total · variance · buckets_sum · violations_listed

Model's actual answer

show the model's full answer
ARR reconciliation as of 2026-09-05 (variance = CRM − billing).

Totals
- CRM (company_arr.csv, 39 companies): 603,581.76
- Billing (chargebee_subscriptions.csv, 39 subs, 1 per company):
  - All statuses: 51,491.48 MRR × 12 = 617,897.76
  - Active only (37 subs; ARR basis): 50,394.94 MRR × 12 = 604,739.28
- Variance on active basis: 603,581.76 − 604,739.28 = −1,157.52
- (All-status basis: 603,581.76 − 617,897.76 = −14,316.00; on that basis the status bucket below nets to zero because CRM matches the cancelled subs exactly.)

Decomposition (sums exactly to −1,157.52)
1. Status mismatch: +13,158.48 — CRM carries ARR for 2 subscriptions billing shows cancelled (4,905.24 + 8,253.24).
2. Rounding: +36.00 — two CRM values rounded to the nearest 100 (16.00 + 20.00).
3. Missing records: −11,952.00 — CRM-only C-0D5BBE3A +16,497.24; billing-only C-21629AA4 −28,449.24.
4. Other: −2,400.00 — C-0F7269D7 active-match variance (24,396.00 − 26,796.00).
Check: 13,158.48 + 36.00 − 11,952.00 − 2,400.00 = −1,157.52 ✓

Mismatched accounts (all other matched accounts reconcile to 0.00)
| Account | Sub | Issue | Delta (CRM−bill) | Suggested owner |
|---|---|---|---|---|
| C-0C8323BF | SUB-000E | Cancelled in billing; CRM still shows 4,905.24 | status | RevOps (CRM hygiene) |
| C-0DC4FB8C | SUB-000F | Cancelled in billing; CRM still shows 8,253.24 | status | RevOps (CRM hygiene) |
| C-0D5BBE3A | — | In CRM (16,497.24), no subscription record | +16,497.24 | Deal Desk / Billing ops |
| C-21629AA4 | SUB-0004 | Active sub (28,449.24), no CRM company record | −28,449.24 | RevOps (CRM data) |
| C-0F7269D7 | SUB-0006 | Active; billing 26,796.00 vs CRM 24,396.00 | −2,400.00 | Deal Desk |
| C-0D66DF9E | SUB-0005 | 23,200.00 vs 23,184.00 | +16.00 | RevOps (data quality) |
| C-14D70CE0 | SUB-0008 | 18,200.00 vs 18,180.00 | +20.00 | RevOps (data quality) |

No owner/rep field exists in either file, so owners are suggested by function, not from data.

Term-rule violations (term ≠ 12 months without cf_agreement_end_date)
- SUB-0002, C-1794A52C, term 24, cf_agreement_end_date empty — VIOLATION
- SUB-0019, C-22170CA1, term 36, cf_agreement_end_date empty — VIOLATION

Compliant non-12-month subs: SUB-000C (C-0DB48281, 24 mo, 2027-11-30), SUB-001A (C-0FC4DBB8, 36 mo, 2027-11-30).

Note: C-1794A52C and C-22170CA1 match CRM exactly on ARR, so these violations are data-completeness only, not ARR variance.
kvm-month-over-monthPASS 1.00
Plain-English question: How did the four core KVMs move in August versus July?
Prompt — what the model saw
Report the four core KVMs for 2026-08 against 2026-07: giving rate, redemptions per user, 1:1 meetings engagement, pulse check engagement. For each: value, prior value, absolute and relative change, direction. Name the KVM with the largest relative move and the segment (plan_tier or size_band) driving it if the data supports that; otherwise say the data does not. Under 200 words plus one table.
cost $0.0335 · 54s · in 14,083 / out 6,006 tok
✓ four_kvms✓ biggest_mover_named✓ direction_correct✓ segment_named✓ word_cap

✓ Correct Answer (ground truth)

Plain-English answer: Giving rate 60.27% (up 0.07%), redemptions per user 1.7302 (up 0.01%), with 1:1 meetings and pulse engagement each up fractionally — direction reported for all four.
The answer in one line: n_company_rows_per_month = 30; aggregation = simple mean of company-level values per month
n_company_rows_per_month30
aggregationsimple mean of company-level values per month
biggest_moverpulse_engagement
driving_segmentsize_band=enterprise; 2026-07=0.55; 2026-08=0.2743; rel_change_pct=-50.13
kvms{"giving_rate": {"2026-07": 0.6023, "2026-08": 0.6027, "abs_change": 0.0004, "rel_change_pct": 0.07, "direction": "up"}, "redemptions_per_user": {"2026-07": 1.73, "2026-08": 1.7302, "abs_change": 0.00
segment_detail{"smb": {"2026-07": 0.6588, "2026-08": 0.6573, "rel_change_pct": -0.23}, "mid_market": {"2026-07": 0.593, "2026-08": 0.5942, "rel_change_pct": 0.2}, "enterprise": {"2026-07": 0.55, "2026-08": 0.2743,
graded checksfour_kvms · biggest_mover_named · direction_correct · segment_named · word_cap

Model's actual answer

show the model's full answer
**Four core KVMs, 2026-08 vs 2026-07** (simple mean across all 30 companies per month; all rows are tier_three).

So-what: three KVMs are essentially flat (|Δ| ≤ 0.1%); pulse check engagement fell 15.3%, entirely because every enterprise company halved.

| KVM | 2026-07 | 2026-08 | Abs Δ | Rel Δ | Direction |
|---|---|---|---|---|---|
| Giving rate | 0.6023 | 0.6027 | +0.0004 | +0.07% | up |
| Redemptions per user | 1.7300 | 1.7302 | +0.0002 | +0.01% | up |
| 1:1 meetings engagement | 0.4469 | 0.4472 | +0.0003 | +0.06% | up |
| Pulse check engagement | 0.6006 | 0.5086 | −0.0920 | −15.31% | down |

Arithmetic: rel Δ = (Aug − Jul) / Jul (e.g., pulse: (0.5086 − 0.6006)/0.6006 = −15.31%).

**Largest relative move:** pulse check engagement (−15.31%). The driving segment is size_band = enterprise: 0.5500 → 0.2743 (−0.2757, −50.13%). All 10 enterprise companies (C-0B2895EF, C-0B2213A9, C-0D6CC8E3, C-0D0B047C, C-0D3278C7, C-0FCCD2DF, C-0F6C0F34, C-8C2E8F00, C-0B827671, C-0BA71F12) dropped uniformly (−0.263 to −0.287 each), while smb (−0.22%) and mid_market (+0.21%) were flat. plan_tier cannot drive anything: every company is tier_three, so no plan-tier contrast exists in the data.
rewards-ytd-sectionPASS 1.00
Plain-English question: Produce the redemption section of the weekly rewards report through August.
Prompt — what the model saw
Produce the redemption section of the weekly rewards report through the last completed month (2026-08): redemption count, spend, unique redeemers, redemptions per redeemer, provider mix as percent of spend, top 5 countries by redemptions. Say which month is the last completed month. Provider shares must sum to 100.
cost $0.1669 · 272s · in 45,568 / out 45,328 tok
✓ count✓ spend✓ unique_redeemers✓ month_named✓ provider_mix✓ top_country

✓ Correct Answer (ground truth)

Plain-English answer: 378 redemptions, $27,846 spend, 235 unique redeemers at 1.61 each, custom rewards at 39% of spend, US leading all countries.
The answer in one line: redemption_count = 378; unique_redeemers = 235; redemptions_per_redeemer = 1.61; total_spend_usd = 27846.00
redemption_count378
redemptions_per_redeemer1.61
unique_redeemers235
provider_mix_pct_of_spendTangoCard=11.6; Tremendous=30.54; custom=39.05; Snappy=18.81
top5_countries_by_countcountry=US; count=244 | country=CA; count=24 | country=AU; count=21 | country=GB; count=17 | country=NL; count=17
total_spend_usd27846.00
graded checkscount · spend · unique_redeemers · month_named · provider_mix · top_country

Model's actual answer

show the model's full answer
Redemption section — through 2026-08 (the last completed month; data spans 2026-01-01 to 2026-08-31, so all 377 rows qualify)

Headline
- Redemptions: 377
- Spend: $27,821.00
- Unique redeemers: 235
- Redemptions per redeemer: 377 / 235 = 1.60

Provider mix (% of spend)
| Provider | Spend (USD) | Share |
|---|---|---|
| custom | 10,873.00 | 39.1% |
| Tremendous | 8,480.00 | 30.5% |
| Snappy | 5,238.00 | 18.8% |
| TangoCard | 3,230.00 | 11.6% |
| Total | 27,821.00 | 100.0% |

Check: 10,873 + 8,480 + 5,238 + 3,230 = 27,821. Shares e.g. 10,873 / 27,821 = 39.08% → 39.1%. Shares sum to 100.0.

Top 5 countries by redemptions
| Country | Redemptions | Spend (USD) |
|---|---|---|
| US | 243 | 18,522.00 |
| CA | 24 | 2,286.00 |
| AU | 21 | 1,606.00 |
| GB | 17 | 944.00 |
| NL | 17 | 1,122.00 |

Note: GB and NL tie at 17 redemptions; NL ranks ahead if ordered by spend within the tie (NL $1,122 vs GB $944). Next countries below the top 5: SG 12, DE/FR/CH 9 each.
customer-success 5 tests
churn-save-eligibilityPASS 1.00
Plain-English question: Which at-risk accounts actually qualify for a churn-save offer, and which play fits each?
Prompt — what the model saw
Which at-risk accounts qualify for a churn-save offer under the documented eligibility rules, what amount is at stake per account and in total, and which play fits each (usage revival, executive touch, commercial concession)? Cite the signal that justifies each play. List accounts that look at risk but do not qualify and why.
cost $0.0428 · 100s · in 6,450 / out 12,184 tok
✓ eligible_set✓ total_at_stake✓ plays_cited✓ noneligible_named✓ rules_applied

✓ Correct Answer (ground truth)

Plain-English answer: Only accounts meeting all three rules qualify — e.g. C-0F6C0F34 at $49,707 (executive touch) and C-0B827671 at $25,365 (usage revival) — with amount and play per account.
The answer in one line: n_accounts = 30; snapshot_date = 2026-09-05
n_accounts30
eligibleaccount_alias=C-0F6C0F34; amount_at_stake=49707.00; play=executive touch; justifying_signal=champion_active is false - no executive sponsor engaged | account_alias=C-0B827671; amount_at_stake=25365.00; play=usage revival; justifying_signal=usage_trend_3m=declining over the last 3 months | account_alias=C-0B360C78; amount_at_stake=35748.00; play=commercial concession; justifying_signal=usage stable/growing with seat utilization 75% - risk is commercial, not adoption | account_alias=C-0B0F1BAB; amount_at_stake=5494.00; play=executive touch; justifying_signal=champion_active is false - no executive sponsor engaged | account_alias=C-0CA21961; amount_at_stake=16829.00; play=usage revival; justifying_signal=seat utilization 26% is below 50% | account_alias=C-0E9C27D1; amount_at_stake=41235.00; play=commercial concession; justifying_signal=usage stable/growing with seat utilization 85% - risk is commercial, not adoption … (+2 more)
non_eligible_at_riskaccount_alias=C-0BC71BDD; health_score=55 | account_alias=C-0BA71F12; health_score=52 | account_alias=C-0F6694C3; health_score=43 | account_alias=C-0BE96399; health_score=54 | account_alias=C-0F876796; health_score=47 | account_alias=C-0FCCD2DF; health_score=43 … (+1 more)
ruleshealth_score < 60, churn_save_eligible_amount > 0, renewal within 120 days of snapshot
snapshot_date2026-09-05
total_amount_at_stake224601.00
graded checkseligible_set · total_at_stake · plays_cited · noneligible_named · rules_applied

Model's actual answer

show the model's full answer
Eligibility screen (R1 AND R2 AND R3), snapshot 2026-09-05, cutoff = 2026-09-05 + 120d = 2027-01-03. 15 accounts fail R1 (health < 60); 8 of those also pass R2 and R3.

QUALIFYING ACCOUNTS (8), amount at stake = churn_save_eligible_amount:

1. C-0F6C0F34 — health 51, $49,707 at stake, renews 2026-10-03 (28d)
   Play: executive touch. Signal: champion_active = false while usage is growing (78% seat util) — adoption is not the problem, the missing sponsor is.

2. C-0B827671 — health 56, $25,365, renews 2026-11-14 (70d)
   Play: usage revival. Signal: usage_trend_3m = declining with 113/202 = 55.9% seat util, but champion_active = true to drive it internally.

3. C-0B360C78 — health 57, $35,748, renews 2026-10-28 (53d)
   Play: commercial concession. Signal: usage growing at 246/327 = 75.2% util with an active champion — usage and relationship look healthy, so the health-57 risk is best addressed on value/commercial terms ahead of a 53-day renewal.

4. C-0B0F1BAB — health 38, $5,494, renews 2026-09-23 (18d)
   Play: executive touch. Signal: champion_active = false with the worst health in the set (38) and only 18 days to renewal; no internal sponsor to revive usage (flat, 238/363 = 65.6% util).

5. C-0CA21961 — health 58, $16,829, renews 2026-12-28 (114d)
   Play: usage revival. Signal: flat usage at only 84/325 = 25.8% seat util with champion_active = true — deep non-use is the churn signal, and there is a champion to act on it.

6. C-0E9C27D1 — health 39, $41,235, renews 2026-09-24 (19d)
   Play: commercial concession. Signal: flat but heavily used (134/157 = 85.4% util) with an active champion — engagement is strong, so a health score of 39 at 19 days from renewal points to a price/value dispute, not adoption or relationship.

7. C-0CEF69FD — health 53, $32,621, renews 2026-11-21 (77d)
   Play: executive touch. Signal: champion_active = false despite growing usage (97/136 = 71.3% util) — adoption is fine, sponsorship is missing.

8. C-0D3278C7 — health 54, $17,602, renews 2026-11-12 (68d)
   Play: usage revival. Signal: usage_trend_3m = declining at 126/380 = 33.2% util with champion_active = true.

Totals:
- Usage revival (3): 25,365 + 16,829 + 17,602 = $59,796
- Executive touch (3): 49,707 + 5,494 + 32,621 = $87,822
- Commercial concession (2): 35,748 + 41,235 = $76,983
- Grand total at stake: 49,707 + 25,365 + 35,748 + 5,494 + 16,829 + 41,235 + 32,621 + 17,602 = $224,601
- Underlying ARR of these 8 accounts: $454,380.

AT RISK (health < 60) BUT DO NOT QUALIFY (7):

- C-0BC71BDD — health 55, renews 2026-10-27 (52d): fails R2, churn_save_eligible_amount = 0.
- C-0BA71F12 — health 52, amount $6,824, renews 2027-04-11 (218d): fails R3, renewal beyond 120 days.
- C-0F6694C3 — health 43, renews 2027-03-21 (197d): fails R2 (amount = 0) and R3 (>120 days).
- C-0BE96399 — health 54, renews 2026-10-29 (54d): fails R2, amount = 0.
- C-0F876796 — health 47, amount $19,958, renews 2027-02-06 (154d): fails R3, renewal beyond 120 days.
- C-0FCCD2DF — health 43, renews 2027-04-23 (230d): fails R2 (amount = 0) and R3 (>120 days).
- C-10A56B0F — health 54, renews 2026-12-12 (98d): fails R2, amount = 0.

Note: the remaining 15 accounts in the file have health >= 60 and fail R1, so they are not considered at risk.

Caveat: the provided files document only eligibility rules (R1-R3). No documented play-assignment rule exists in the data; the play mapping above is analyst judgment using the three usage signals (usage_trend_3m, seat utilization, champion_active): no active champion -> executive touch; declining or low-utilization usage with an active champion -> usage revival; healthy usage plus active champion -> commercial concession.
expansion-kitPASS 1.00
Plain-English question: Build the expansion kit for the customer who just replied to our outreach.
Prompt — what the model saw
A customer replied to outreach (reply included). Build the expansion kit: seat coverage (licensed over headcount), two lines on usage health, headroom in seats and ARR at the current per-seat rate, who replied and whether they can buy, the right buyer from our contacts, and a reply email under 150 words that moves toward the right person without being pushy and cites one real usage fact.
cost $0.0093 · 23s · in 1,143 / out 2,473 tok
✓ coverage_pct✓ headroom_arr✓ right_buyer✓ replier_not_buyer✓ usage_fact✓ email_under_150

✓ Correct Answer (ground truth)

Plain-English answer: 37.5% seat coverage, 250 headroom seats worth $15,000 ARR at $60 a seat, usage up 43.2% in six months — and the replier cannot buy; Dana R. (VP People) is the right buyer.
The answer in one line: headroom_seats = 250; coverage_pct = 37.5; company_alias = C-0DDFC9A7 (+3 more below)
headroom_seats250
coverage_pct37.5
company_aliasC-0DDFC9A7
headroom_arr15000.00
per_seat_annual_rate60.00
replierMaria S.
replier_can_buy0
right_buyername=Dana R.; title=VP People
usage_factsactive_users_2026_08=126; active_users_2026_03=88; growth_pct_6m=43.2; licensed_seat_utilization_pct=84
usage_health_linesMonthly active users grew from 88 (2026-03) to 126 (2026-08), +43.2% over 6 months., Last month 126 of 150 licensed seats were active (84.0% utilization), approaching the seat cap.
graded checkscoverage_pct · headroom_arr · right_buyer · replier_not_buyer · usage_fact · email_under_150

Model's actual answer

show the model's full answer
EXPANSION KIT — C-0DDFC9A7

SEAT COVERAGE
- Licensed seats 150 vs headcount 400 → 150/400 = 37.5% of the company is licensed. 250 employees (62.5%) have no seat.

USAGE HEALTH (two lines)
- MAU rose every month, Mar→Aug 2026: 88 → 95 → 102 → 110 → 118 → 126 (+38 users, +43.2%; monthly gains +7, +7, +8, +8, +8; no dips).
- Aug MAU of 126 is 84.0% of the 150 licensed seats (126/150), so usage is nearing license capacity.

HEADROOM AT CURRENT PER-SEAT RATE
- Current per-seat rate: $9,000 / 150 seats = $60/seat.
- Headroom vs headcount: 400 − 150 = 250 seats → 250 × $60 = $15,000 ARR.
- Headroom vs current usage (tighter read): 150 − 126 = 24 seats → 24 × $60 = $1,440 ARR; at Aug's +8 MAU/month, licenses saturate in ~3 months (≈Nov 2026), so the expansion trigger is usage, not just headcount.

WHO REPLIED / CAN THEY BUY
- Maria S., People Operations Coordinator (last engaged 2026-09-02, matches the reply). She cannot buy: she states purchasing, budget, and seat expansion sit with Dana R. She offered an introduction — use her as the path, not the target.

RIGHT BUYER
- Dana R., VP People (last engaged 2026-05-18, ~3.5 months stale). She is named in the reply as the budget/seat owner, and Maria says Dana has been asking about usage numbers — warm, relevant entry point. Sam K. (Office Manager, last engaged 2025-11-03) is not a buyer.

REPLY EMAIL (83 words)
Subject: RE: Growing your team's recognition program

Hi Maria,

Glad to hear the team is enjoying Bonusly — the numbers back it up: monthly active users climbed from 88 in March to 126 in August, up about 43% in five months.

Thanks for pointing me to Dana. An introduction would be great whenever you have a moment — no rush. I'll keep it brief: share the usage trend, walk through what seat expansion would look like, and leave the decision in her hands.

Appreciate you making this easy.

Best,
Cole

DATA GAPS
- Contacts file has no email addresses, seniority flags, or relationship roles beyond titles; buyer identification relies on titles plus Maria's reply.
- No renewal date, contract terms, or prior pricing history provided; headroom assumes flat $60/seat pricing.
onboarding-statusPASS 1.00
Plain-English question: What is done and what is missing in this customer's onboarding before the mid-onboarding call?
Prompt — what the model saw
Prep the CSM for the mid-onboarding call: what is complete (integrations connected, allowance set, admins added, first recognition given, first redemption), what is not, early engagement signals, and the three things to cover on the call. Mark nothing complete without a data field that shows it.
cost $0.0163 · 36s · in 1,222 / out 4,806 tok
✓ complete_items✓ incomplete_items✓ three_items✓ grounded

✓ Correct Answer (ground truth)

Plain-English answer: 25 days live: Slack connected, allowance set, admins added — but the HRIS integration is not connected, with early engagement signals and three call priorities.
The answer in one line: days_live = 25; company_alias = C-0D284E42
days_live25
call_itemsConnect the HRIS integration (integration_hris is empty) to automate user provisioning and offboarding., Get to first redemption (first_redemption_at is empty): walk admins through the rewards catalog and confirm allowance funding., Build the giving habit: daily active givers averaged 4.3 in week 1 vs 13.0 last week - agree on manager nudges and a recognition cadence.
checklist{"slack_integration_connected": {"complete": true, "field": "integration_slack", "value": "2026-08-12"}, "hris_integration_connected": {"complete": false, "field": "integration_hris", "value": ""}, "a
company_aliasC-0D284E42
engagement_signalsavg_active_givers_week1=4.3; avg_active_givers_last_week=13; latest_active_givers=15; days_with_activity=25
graded checkscomplete_items · incomplete_items · three_items · grounded

Model's actual answer

show the model's full answer
ONBOARDING PREP — C-0D284E42
Signup 2026-08-11. Usage data runs 2026-08-11 through 2026-09-04 (25 days of data; day 24 post-signup at the last row).

WHAT IS COMPLETE (each backed by a field in onboarding_account.csv)
- Slack integration: complete. integration_slack = 2026-08-12, +1 day after signup.
- Allowance set: complete. allowance_set = 2026-08-13, +2 days after signup.
- First recognition given: complete. first_recognition_at = 2026-08-15 14:22, +4 days after signup.
- Admins added: 2 per admins_added = 2. The field confirms two admins; no target count is given, so "complete" can't be judged — only the count is evidenced.

NOT COMPLETE / NO EVIDENCE
- HRIS integration: not complete. integration_hris is blank.
- First redemption: not complete. first_redemption_at is blank. That is 24 days after signup and 20 days after the first recognition (2026-08-15 to 2026-09-04) with no redemption recorded.

EARLY ENGAGEMENT SIGNALS (onboarding_usage.csv, active_givers)
- Net growth: 3 (2026-08-11) to 15 (2026-09-04) = +12 givers, 15/3 = 5.0x over 24 days.
- Weekly averages: Aug 11–17 = 30/7 = 4.3; Aug 18–24 = 53/7 = 7.6; Aug 25–31 = 75/7 = 10.7; Sep 1–4 = 56/4 = 14.0. Every period higher than the last.
- Rolling week: Aug 22–28 = 66/7 = 9.4 vs Aug 29–Sep 4 = 91/7 = 13.0, +38% week over week.
- Dips (e.g., Aug 16: 4 after 5; Aug 31: 11 after 13) are one-day and recover; no sustained decline in the 25-day series.
- Caution: last two days flat at 15 (Sep 3, Sep 4) — growth paused at the very end of the window.
- Risk signal: recognition started on day 4 and giver counts keep growing, yet redemptions are zero. Value is being given but nothing is being redeemed 20 days later.

THREE THINGS TO COVER ON THE CALL
1. Unblock the first redemption. Evidence: first_redemption_at blank through day 24 despite 15 active givers and a first recognition on 2026-08-15. Walk the rewards catalog and redemption path with the admins; verify the allowance set 2026-08-13 is visible and usable where recognition happens (Slack, connected 2026-08-12); identify the blocker (catalog, budget visibility, awareness).
2. Close the HRIS gap. Evidence: integration_hris blank while Slack was connected on day 1. Cover what HRIS sync unlocks (employee/department sync, new-hire and off-boarding hygiene) and leave the call with an owner and date for setup.
3. Admin depth and the end-of-window plateau. Evidence: admins_added = 2 against active_givers growth 3 to 15, and the last two days flat at 15. Discuss adding admins/champions to carry momentum past the initial 15 givers and keep the Sep 3–4 plateau from becoming the new baseline.

MISSING DATA (cannot be answered from what was provided)
- No employee or seat count, so giver penetration is unknown; 15 active givers could be a small share or the whole company.
- No fields on redemption attempts, catalog views, or budget visibility, so the reason for zero redemptions cannot be derived from data — it must be asked on the call.
- No target admin count, so admins_added = 2 is a count, not a completion status.
renewal-risk-conflicting-datesPASS 1.00
Plain-English question: Two systems disagree on renewal dates — which do you trust per account, and who is actually renewing inside 90 days?
Prompt — what the model saw
Produce a 90-day renewal risk brief. Two systems hold renewal dates and disagree on some accounts; decide which to trust per account and say why (multi-year contracts are known to be wrong in ChurnZero). For every renewal: company, CSM, ARR, date used, seat utilization, 3-month usage trend, risk rating with one sentence of evidence. Flag every disagreement. Close with total ARR renewing and ARR at risk.
cost $0.1054 · 240s · in 14,437 / out 31,338 tok
✓ total_renewing✓ arr_at_risk✓ disagreements_flagged✓ trust_rule

✓ Correct Answer (ground truth)

Plain-English answer: Chargebee wins per account (multi-year contracts are known wrong in ChurnZero), with each trusted date, the reason, and a 90-day-window flag.
The answer in one line: n_accounts = 20; n_disagreements = 5; snapshot_date = 2026-09-05 (+2 more below)
n_accounts20
n_disagreements5
accounts20 items — e.g. account_alias=C-0B144C78; csm=Cole Ingram; arr=30899.00; trusted_renewal_date=2026-11-02; trusted_source_why=systems agree (annual term); in_90d_window=True; dates_disagree=False; seat_utilization_pct=75.4; usage_3m_ratio=1.03; risk=low; evidence=3-month usage ratio 1.03 (last3 avg 103 vs prior3 100), seat utilization 75% | account_alias=C-0B20DB64; csm=Dana Mercer; arr=21770.00; trusted_renewal_date=2026-10-07; trusted_source_why=systems agree (annual term); in_90d_window=True; dates_disagree=False; seat_utilization_pct=56.6; usage_3m_ratio=1; risk=medium; evidence=3-month usage ratio 1.00 (last3 avg 295 vs prior3 295), seat utilization 57% | account_alias=C-0B344485; csm=Elena Sinclair; arr=64384.00; trusted_renewal_date=2026-11-16; trusted_source_why=systems agree (annual term); in_90d_window=True; dates_disagree=False; seat_utilization_pct=78; usage_3m_ratio=1.04; risk=low; evidence=3-month usage ratio 1.04 (last3 avg 241 vs prior3 231), seat utilization 78% | account_alias=C-0B369871; csm=Dana Mercer; arr=85128.00; trusted_renewal_date=2026-10-29; trusted_source_why=systems agree (annual term); in_90d_window=True; dates_disagree=False; seat_utilization_pct=75.1; usage_3m_ratio=1.04; risk=low; evidence=3-month usage ratio 1.04 (last3 avg 330 vs prior3 316), seat utilization 75%
arr_at_risk_90d359409.00
disagreementsaccount_alias=C-0B7D2C30; cz_renewal_date=2026-09-10; cb_renewal_date=2026-09-15; is_multi_year=True; trusted=2026-09-15 | account_alias=C-0BBE3E60; cz_renewal_date=2027-09-26; cb_renewal_date=2026-09-26; is_multi_year=True; trusted=2026-09-26 | account_alias=C-0BCDB8C2; cz_renewal_date=2027-09-18; cb_renewal_date=2026-09-18; is_multi_year=True; trusted=2026-09-18 | account_alias=C-0D2AB865; cz_renewal_date=2026-09-10; cb_renewal_date=2026-09-22; is_multi_year=True; trusted=2026-09-22 | account_alias=C-0F5D2323; cz_renewal_date=2026-09-10; cb_renewal_date=2026-09-29; is_multi_year=True; trusted=2026-09-29
snapshot_date2026-09-05
total_arr_renewing_90d1048715.00
trust_rulemulti-year contracts: Chargebee is authoritative (ChurnZero known wrong); otherwise systems agree or Chargebee wins
window2026-09-05 to 2026-12-04
graded checkstotal_renewing · arr_at_risk · disagreements_flagged · trust_rule

Model's actual answer

show the model's full answer
90-DAY RENEWAL RISK BRIEF — data as of Oct 19, 2026; window = Oct 19, 2026 → Jan 17, 2027 (+90 days)

So-what: The next 90 days carry 10 renewals / $516,117, and nearly all are healthy (one Medium blip, $40,628). The real exposure is behind us: 10 accounts (all 5 date-disagreements plus 5 agreed) had renewal dates before Oct 19 with no outcome recorded in these files, and 6 of them are High risk ($359,409) on collapsing usage. All 5 system disagreements are multi-year accounts and resolve to Chargebee.

DATE-TRUST RULE (applied per account)
Known defect: multi-year contracts are wrong in ChurnZero. All 5 disagreements are the only multi-year rows (term 24/36 mo); the other 15 accounts (12-mo) match exactly in both systems. So: multi-year → trust Chargebee; 12-mo → either (they agree). Corroboration: ChurnZero shows three different accounts all stamped 2026-09-10 (a placeholder-looking cluster) where Chargebee shows staggered dates (Sep 15/22/29).

DISAGREEMENTS FLAGGED (5; all multi-year; all resolved to Chargebee)
- C-0B7D2C30: CZ 2026-09-10 vs CB 2026-09-15 (36 mo) → used 2026-09-15. CZ off by +5 days, sits in the 09-10 placeholder cluster.
- C-0D2AB865: CZ 2026-09-10 vs CB 2026-09-22 (24 mo) → used 2026-09-22. Same cluster.
- C-0F5D2323: CZ 2026-09-10 vs CB 2026-09-29 (24 mo) → used 2026-09-29. Same cluster.
- C-0BCDB8C2: CZ 2027-09-18 vs CB 2026-09-18 (36 mo) → used 2026-09-18. CZ off by exactly +365 days; trusting CZ would hide this as FY27.
- C-0BBE3E60: CZ 2027-09-26 vs CB 2026-09-26 (24 mo) → used 2026-09-26. CZ off by exactly +365 days, same pattern.
Impact: ChurnZero misstates $85,420 ($54,427 + $30,993) of already-due renewals as 2027 renewals, and shifts 3 more by 5–19 days.

Risk rules used: High = 3-mo usage decline ≥10% OR seat utilization <40%; Medium = not High and (3-mo decline OR utilization <70%); Low = growth AND utilization ≥70%. Seat utilization = seats_used / seats. 3-month trend = (Aug 2026 − May 2026) / May 2026 active users.

SECTION A — RENEWALS INSIDE THE 90-DAY WINDOW (10)
C-0BC34584 | Cole Ingram | $16,740 | 2026-10-22 (both agree) | util 66.2% (327/494) | +2.9% (103→106) | Medium — usage growing but only 66% of seats used and active users (106) far below seats.
C-0B7A7546 | Elena Sinclair | $35,062 | 2026-10-25 (both agree) | util 88.8% (182/205) | +3.3% (61→63) | Low — high utilization, stable growth.
C-0B369871 | Dana Mercer | $85,128 | 2026-10-29 (both agree) | util 75.1% (317/422) | +4.4% (319→333) | Low — growing 15% YoY on top utilization.
C-0B144C78 | Cole Ingram | $30,899 | 2026-11-02 (both agree) | util 75.4% (169/224) | +7.1% (99→106) | Low — fastest grower in window, healthy utilization.
C-0FC4DBB8 | Elena Sinclair | $94,732 | 2026-11-05 (both agree) | util 76.7% (356/464) | +4.3% (185→193) | Low — steady growth, mid-70s utilization; largest renewal in window.
C-0D5BBE3A | Dana Mercer | $39,740 | 2026-11-09 (both agree) | util 83.3% (85/102) | +4.6% (87→91) | Low — near-saturated seats (89% active/seat), growing.
C-0FB9D5AF | Cole Ingram | $63,158 | 2026-11-13 (both agree) | util 72.4% (144/199) | +4.8% (168→176) | Low — 88% of seats active, consistent growth.
C-0B344485 | Elena Sinclair | $64,384 | 2026-11-16 (both agree) | util 78.0% (224/287) | +3.8% (235→244) | Low — steady growth, strong utilization.
C-0CB2C1B4 | Dana Mercer | $40,628 | 2026-11-20 (both agree) | util 81.6% (386/473) | -2.0% (50→49) | Medium — only decliner in window; active users (49) equal ~10% of seats despite 81.6% seat utilization, so the seats_used figure may be unreliable.
C-22170CA1 | Cole Ingram | $45,646 | 2026-11-24 (both agree) | util 85.4% (251/294) | +2.1% (143→146) | Low — high utilization, mild growth.
Window total: 16,740+35,062+85,128+30,899+94,732+39,740+63,158+64,384+40,628+45,646 = $516,117.
Window at risk (Medium+High): 16,740+40,628 = $57,368 (11.1% of $516,117). High risk: $0.

SECTION B — RENEWAL DATES ALREADY PASSED BEFORE THE WINDOW (10) — outcome not in the data
Missing data, stated explicitly: none of the three files records renewal outcome (renewed/churned/expanded), so these are pre-renewal health ratings; treat as open exposure until confirmed.
C-0B7D2C30 | Dana Mercer | $65,901 | 2026-09-15 (Chargebee; disputed) | util 57.6% (274/476) | -21.5% (107→84) | High — usage down 21.5% in 3 mo and 46% YoY on 58% seat utilization.
C-0BCDB8C2 | Cole Ingram | $54,427 | 2026-09-18 (Chargebee; disputed) | util 54.7% (232/424) | -19.1% (136→110) | High — steady ~20% quarterly decline, half of seats unused; date hidden as 2027 by ChurnZero.
C-0D2AB865 | Elena Sinclair | $38,022 | 2026-09-22 (Chargebee; disputed) | util 61.4% (250/407) | -20.4% (137→109) | High — 20% 3-mo decline, ~45% YoY decline.
C-0BBE3E60 | Dana Mercer | $30,993 | 2026-09-26 (Chargebee; disputed) | util 64.9% (74/114) | -19.5% (41→33) | High — smallest account, worst relative decline (-48% YoY); CZ date off by a year.
C-0F5D2323 | Cole Ingram | $90,647 | 2026-09-29 (Chargebee; disputed) | util 28.5% (111/390) | -10.0% (20→18) | High — only 18 active users against 390 seats (4.6%); lowest engagement of any high-ARR account.
C-0EC6999D | Elena Sinclair | $79,419 | 2026-10-03 (both agree) | util 27.7% (31/112) | +7.1% (14→15) | High — utilization under 30% and only ~15 active users despite flat-to-up usage; biggest renewal just past due.
C-0B20DB64 | Dana Mercer | $21,770 | 2026-10-07 (both agree) | util 56.6% (214/378) | -0.7% (296→294) | Medium — heavy daily usage (294 active) but only 57% of seats used; flat, not declining.
C-0BBC4E7A | Cole Ingram | $56,374 | 2026-10-10 (both agree) | util 67.7% (228/337) | -2.1% (142→139) | Medium — flat usage with mild seat underuse.
C-0FD551AB | Elena Sinclair | $48,815 | 2026-10-14 (both agree) | util 55.9% (210/376) | +0.8% (125→126) | Medium — stable usage but only 56% seat utilization.
C-0F9F8F13 | Dana Mercer | $46,230 | 2026-10-18 (both agree) | util 56.5% (199/352) | 0.0% (182→182) | Medium — perfectly flat usage, ~44% of seats unused; due date was yesterday.
Past-due total: 65,901+54,427+38,022+30,993+90,647+79,419+21,770+56,374+48,815+46,230 = $532,598.
Past-due High: 65,901+54,427+38,022+30,993+90,647+79,419 = $359,409. Past-due Medium: 21,770+56,374+48,815+46,230 = $173,189.

DATA-QUALITY NOTES
- seats_used and active_users disagree on several accounts (e.g., C-0BC34584: 327 used vs 106 active; C-0CB2C1B4: 386 vs 49; C-0B20DB64: 214 used vs 294 active). I report seats_used/seats as "seat utilization" per the file and usage trend from active_users; where they conflict, the seats figure is suspect.
- No renewal-outcome, renewal-status, or expansion/churn field exists in these files — Section B ratings are health, not outcomes.
- Usage series ends Aug 2026; no Sep/Oct signal available for the 10 most recent renewals.

CLOSING TOTALS
- ARR renewing inside the next 90 days (Oct 19 – Jan 17, 2027): $516,117. ARR at risk in window: $57,368 (both Medium; $0 High).
- ARR with renewal dates already passed (outcome missing from data): $532,598; at risk: $532,598 (High $359,409 + Medium $173,189).
- Entire file (20 accounts, Sep 10 – Nov 24 dates): $516,117 + $532,598 = $1,048,715; total ARR at risk on the High+Medium definition: $57,368 + $532,598 = $589,966 (56.3%); High-risk only: $359,409 (34.3%).
ticket-theme-synthesisPASS 1.00
Plain-English question: What are the real themes in this quarter's support tickets, ignoring the existing tags?
Prompt — what the model saw
Synthesize the quarter's support tickets into 4 to 7 themes. Do not trust existing tags; read the text. Per theme: name, count, share, distinct accounts, ARR affected, two ticket ids, one-line recommendation. Separate broad patterns from single-account noise and rank by ARR exposure, not volume.
cost $0.0722 · 128s · in 18,405 / out 17,860 tok
✓ theme_count_range✓ counts_verifiable✓ single_account_flagged✓ arr_ranked✓ ticket_ids

✓ Correct Answer (ground truth)

Plain-English answer: 80 tickets in 4-7 themes — points-not-posting leads (20 tickets, 25%, $31,100 ARR affected) — each with counts, distinct accounts, ARR, and a recommendation.
The answer in one line: total_tickets = 80; quarter_window = 2026-06-01 through 2026-08-31
total_tickets80
noteexisting tags are noisy/unreliable; classify by text
quarter_window2026-06-01 through 2026-08-31
ranking_by_arr_exposurehris_provisioning, redemption_failures, billing_dispute, points_not_posting, slack_sync
sample_ticket_ids80 items — e.g. IC-460004, IC-460016, IC-460001, IC-460041
single_account_noise_themebilling_dispute
themes{"points_not_posting": {"count": 20, "share_pct": 25.0, "distinct_accounts": 9, "arr_affected": "31100.00", "example_ticket_ids": ["IC-460001", "IC-460002"], "single_account": false}, "slack_sync": {"
graded checkstheme_count_range · counts_verifiable · single_account_flagged · arr_ranked · ticket_ids

Model's actual answer

show the model's full answer
All 80 tickets classified from body text (June 1 – Aug 29, 2026). Tags are unreliable and were ignored: e.g., IC-460020 and IC-460025 are tagged "billing" but describe points/redemption failures; IC-460077 is tagged "how-to" but is an invoice dispute; HRIS tickets carry tags "urgent", "billing", "question" alike. 24 distinct accounts, total ARR in scope $284,800 (no account appears in more than one theme, so theme ARR sums add without double counting).

THEMES, RANKED BY ARR EXPOSURE

1. HRIS provisioning failures — broad pattern
- Count: 12 (15.0% of 80)
- Distinct accounts: 3 (C-0B2213A9, C-0DDFC9A7, C-0F6C0F34)
- ARR affected: 36,000 + 48,000 + 30,000 = $114,000
- Ticket ids: IC-460059, IC-460062
- Recommendation: Treat as enterprise-blocking defect — sync silently skips new hires (provisioning log shows no errors), so add failure alerts and reconciliation jobs, and manually backfill skipped hires for the three accounts.

2. Gift-card / reward redemption failures (incl. points deducted without fulfillment) — broad pattern
- Count: 18 (22.5% of 80)
- Distinct accounts: 7 (C-0B0F1BAB, C-0B827671, C-0CEF69FD, C-0D9CA315, C-0F876796, C-0FCCD2DF, C-14264ABD)
- ARR affected: 10,300 + 10,700 + 8,900 + 9,600 + 8,700 + 9,600 + 11,000 = $68,800
- Ticket ids: IC-460024, IC-460035
- Recommendation: Fix the checkout/redemption path and auto-refund points when a gift-card order errors out (IC-460024, IC-460026, IC-460037 show deductions with no reward), plus monitor for failed deliveries.

3. Invoice and renewal billing disputes — SINGLE-ACCOUNT NOISE
- Count: 16 (20.0% of 80)
- Distinct accounts: 1 (C-0E9C27D1)
- ARR affected: $52,000
- Ticket ids: IC-460069, IC-460078
- Recommendation: One account drives all 16 tickets (recurring seat-count overbilling, unapproved seats, wrong renewal tier); audit its contract/seat sync, issue credits, and route to a single billing owner — this is an escalation, not a product-wide defect.

4. Points not posting after recognitions — broad pattern
- Count: 20 (25.0% of 80)
- Distinct accounts: 9 (C-0B2895EF, C-0BE96399, C-0BF20542, C-0D0B047C, C-0D284E42, C-0D3278C7, C-0D6CC8E3, C-0DD0626C, C-21FEBCBB)
- ARR affected: 2,900 + 2,700 + 4,500 + 4,500 + 3,400 + 3,500 + 4,200 + 2,500 + 2,900 = $31,100
- Ticket ids: IC-460004, IC-460016
- Recommendation: Highest volume, low dollar exposure — audit the recognition-to-points pipeline (several reports point to weekend batch posting) and backfill unposted points.

5. Slack integration failures (sync toggle resets, re-auth not sticking, slash-command errors) — broad pattern
- Count: 14 (17.5% of 80)
- Distinct accounts: 4 (C-0B843542, C-0BA71F12, C-10A56B0F, C-8C2E8F00)
- ARR affected: 4,400 + 3,900 + 5,400 + 5,200 = $18,900
- Ticket ids: IC-460041, IC-460049
- Recommendation: Harden Slack auth/token persistence and sync settings so toggles and re-auths stick; fix the failing slash command for the four affected accounts.

BROAD PATTERNS vs SINGLE-ACCOUNT NOISE
- Broad patterns (multi-account, recurring symptoms): themes 1, 2, 4, 5 — 64 tickets, $232,800 ARR.
- Single-account noise: theme 3 — all 16 tickets from C-0E9C27D1 (20% of quarter volume from one account); it ranks 3rd by ARR but should not be read as a systemic billing-product issue.
- Minor single-ticket accounts within broad themes: C-21FEBCBB (points), C-8C2E8F00 (Slack), C-0D9CA315 and C-0B0F1BAB (redemption) — one ticket each; watch, don't escalate.
- Concentration note: ARR exposure is dominated by 4 enterprise accounts — C-0E9C27D1 ($52k), C-0DDFC9A7 ($48k), C-0B2213A9 ($36k), C-0F6C0F34 ($30k) = $166,000, i.e. 58% of total ARR in scope ($166,000 / $284,800).

Arithmetic checks: counts 12+18+16+20+14 = 80; shares 15.0+22.5+20.0+25.0+17.5 = 100%; ARR 114,000+68,800+52,000+31,100+18,900 = $284,800; distinct accounts 3+7+1+9+4 = 24 (no overlap).
marketing 5 tests
case-study-matchPASS 1.00
Plain-English question: Which three existing customers are the best case-study matches for this prospect?
Prompt — what the model saw
Given the prospect, rank the three most similar existing customers with public case studies for use as social proof. Similarity on industry, size band, use case, region. Cite the fields that drove each match. Do not name any customer without a case study and never include billing data or contact names.
cost $0.0111 · 30s · in 1,411 / out 3,025 tok
✓ top3✓ decoys_excluded✓ cites_fields✓ no_billing

✓ Correct Answer (ground truth)

Plain-English answer: Three ranked matches on industry, size, use case, and region (e.g. C-11C31562), with the higher-scoring decoy explicitly excluded for a mismatched field.
The answer in one line: prospect_alias = C-82AF3719
constraintsno_billing_data, no_contact_names, only_customers_with_has_case_study_true_named
decoys_excludedcompany_alias=C-D6217CAA; similarity_score=4; reason=no_case_study | company_alias=C-C153868F; similarity_score=4; reason=no_case_study
prospect_aliasC-82AF3719
prospect_fields_usedindustry, size_band, use_case, region
top3company_alias=C-11C31562; similarity_score=3 | company_alias=C-64171065; similarity_score=3 | company_alias=C-A13C193D; similarity_score=3
graded checkstop3 · decoys_excluded · cites_fields · no_billing

Model's actual answer

show the model's full answer
Prospect C-82AF3719: Technology / Mid-Market / employee_recognition / NA-West.

Method: score = number of exact matches across the 4 fields (industry, size_band, use_case, region), computed only over customers with has_case_study=true (8 of 25 rows qualify). No case-study customer matches all 4 fields; the best achievable is 3/4.

RANKING

1. C-64171065 — 3/4
   Matched: industry=Technology, size_band=Mid-Market, use_case=employee_recognition.
   Differs: region=NA-East (prospect NA-West; same NA macro-region).
   Why first: only candidate matching both industry and use case plus size band; same vertical and same program is the strongest social-proof pairing.

2. C-11C31562 — 3/4
   Matched: size_band=Mid-Market, use_case=employee_recognition, region=NA-West.
   Differs: industry=Manufacturing.
   Why second: exact use case, size, and region; industry is the only miss.

3. C-A13C193D — 3/4
   Matched: industry=Technology, size_band=Mid-Market, region=NA-West.
   Differs: use_case=retention (prospect employee_recognition).
   Why third: same industry, size, and region, but the use-case miss weakens program-specific proof versus ranks 1-2.

Tiebreak rule applied (all three tie at 3/4): prefer industry match first, then use_case match, then region match. Under it: C-64171065 (industry+use case) > C-11C31562 (use case, no industry) > C-A13C193D (industry, no use case).

FULL SCORE TABLE (case-study-eligible only)
- C-64171065  Tech/MM/employee_recognition/NA-East     3
- C-11C31562  Mfg/MM/employee_recognition/NA-West      3
- C-A13C193D  Tech/MM/retention/NA-West                3
- C-CD4829A7  Tech/MM/retention/NA-East                2 (industry, size)
- C-9D69D147  FinServ/MM/retention/NA-West             2 (size, region)
- C-FEFCC020  Edu/Enterprise/wellness/NA-East          0
- C-11209339  Mfg/SMB/peer_milestones/APAC             0
- C-0CE315B2  Edu/SMB/wellness/EMEA                    0

Notes
- C-CD4829A7 and C-9D69D147 are the next-best fallbacks at 2/4 if more than three names are needed.
- One record in customers.csv matches the prospect on all four fields exactly but has has_case_study=false; it is excluded per the no-case-study rule and is not named here. Same for all other has_case_study=false rows (including the other Technology/Mid-Market/employee_recognition entries).
- No billing data or contact names appear in the source files; none included.
channel-attribution-mathPASS 1.00
Plain-English question: Compute cost per SQM, cost per SQO, and pipeline per dollar for every channel over six months.
Prompt — what the model saw
Compute channel performance for the trailing 6 months. Paid: spend, SQMs, SQOs, cost per SQM, cost per SQO, SQM-to-SQO rate, pipeline amount, pipeline per dollar. Organic: volume, SQO rate, pipeline. Report a channel with spend and zero SQMs as undefined, not zero. Flag rows where SQO date precedes SQM date. Finish with a reallocation recommendation and your confidence given sample sizes.
cost $0.0313 · 57s · in 10,322 / out 7,491 tok
✓ undefined_channel✓ undefined_not_zero✓ flagged_rows✓ best_channel✓ reallocation

✓ Correct Answer (ground truth)

Plain-English answer: LinkedIn Ads at $960/SQM and $3,000/SQO, organic search at $90,000 pipeline on zero spend, paid social flagged at $18,000 with zero SQMs, and a channel winner named.
The answer in one line: best_pipeline_per_dollar_channel = paid_search
best_pipeline_per_dollar_channelpaid_search
confidence_notewebinars and referral have small SQM samples (<20); treat rates as directional.
flagged_rows_sqo_before_sqmCT-000041, CT-000044
months2026-03, 2026-04, 2026-05, 2026-06, 2026-07, 2026-08
per_channel{"linkedin_ads": {"spend_usd": 24000, "sqms": 25, "sqos": 8, "pipeline_amount": 96000, "cost_per_sqm": 960.0, "cost_per_sqo": 3000.0, "sqm_to_sqo_rate": 0.32, "pipeline_per_dollar": 4.0}, "organic_sea
reallocation_recommendationfrom=paid_social; to=paid_search; rationale=paid_social has spend but zero SQMs (cost per SQM undefined, not zero); paid_search has the highest pipeline per dollar.
undefined_channelspaid_social
graded checksundefined_channel · undefined_not_zero · flagged_rows · best_channel · reallocation

Model's actual answer

show the model's full answer
Trailing 6 months = 2026-03 through 2026-08. All 122 contact rows fall in this window; all spend rows are the 6 months × 4 paid channels.

PAID CHANNELS

| channel | spend | SQMs | SQOs | cost/SQM | cost/SQO | SQM→SQO | pipeline | pipeline/$ |
|---|---|---|---|---|---|---|---|---|
| paid_search | $36,000 | 40 | 18 | $900 | $2,000 | 45.0% | $720,000 | 20.0 |
| linkedin_ads | $24,000 | 25 | 8 | $960 | $3,000 | 32.0% | $96,000 | 4.0 |
| webinars | $9,000 | 12 | 5 | $750 | $1,800 | 41.7% | $60,000 | 6.7 |
| paid_social | $18,000 | 0 | 0 | undefined | undefined | undefined | $0 | undefined |

Arithmetic:
- Spend = 6 months × monthly: paid_search 6×6,000=36,000; linkedin_ads 6×4,000=24,000; paid_social 6×3,000=18,000; webinars 6×1,500=9,000. Total paid spend $87,000.
- paid_search: 36,000/40=900; 36,000/18=2,000; 18/40=45.0%; 18 SQOs × 40,000 = 720,000; 720,000/36,000=20.0.
- linkedin_ads: 24,000/25=960; 24,000/8=3,000; 8/25=32.0%; 8 × 12,000 = 96,000; 96,000/24,000=4.0.
- webinars: 9,000/12=750; 9,000/5=1,800; 5/12=41.7%; 5 × 12,000 = 60,000; 60,000/9,000=6.67.
- paid_social: spend present, SQMs=0 → all ratios undefined (per instruction), not zero.

ORGANIC / UNPAID CHANNELS

- organic_search: volume 30 SQMs, SQOs 10, SQO rate 10/30 = 33.3%, pipeline 10 × 9,000 = $90,000.
- referral (unpaid, reported separately): volume 15 SQMs, SQOs 6, SQO rate 6/15 = 40.0%, pipeline 6 × 8,000 = $48,000.

DATA QUALITY FLAGS — SQO date precedes SQM date (2 rows, both linkedin_ads):
- CT-000044: sqm_date 2026-07-23, sqo_date 2026-07-18, pipeline 12,000
- CT-000041: sqm_date 2026-06-14, sqo_date 2026-06-09, pipeline 12,000

Both are currently counted in linkedin_ads SQO/pipeline above. If excluded as unreliable: linkedin_ads SQOs = 6, cost/SQO = 24,000/6 = $4,000, pipeline = 72,000, pipeline/$ = 3.0.

REALLOCATION RECOMMENDATION

1. paid_search is the clear winner: 20.0 pipeline per dollar vs 4.0–6.7 for other paid channels, best conversion (45%) and second-lowest cost per SQM. Fund it first; add budget up to the point its CAC economics hold.
2. paid_social ($18,000, 6 months, zero tracked SQMs): undefined performance, not proven failure. Action is audit-then-decide: verify tracking/attribution first; if tracking is confirmed correct, reallocate its budget to paid_search.
3. linkedin_ads is the weakest measured channel (32% SQM→SQO, pipeline/$ 4.0, and 2 of its 8 SQOs have impossible date ordering). Pause or cut it and shift toward paid_search; revisit only with better conversion.
4. webinars looks efficient (lowest cost/SQM at $750, 41.7% rate, 6.7 pipeline/$) but rests on 12 SQMs and 5 SQOs. Incrementally test +50–100% spend rather than assuming linear returns.
5. organic_search (33.3%) and referral (40.0%, highest SQO rate of any channel) need no spend decision from this data; referral's conversion rate argues for investing effort (not budget) in referrals.

CONFIDENCE

Moderate for the headline ranking paid_search > webinars ≈ linkedin_ads > paid_social-undefined: paid_search rests on 40 SQMs/18 SQOs, organic_search 30/10. Low for webinars (12 SQMs/5 SQOs) and for linkedin_ads (25 SQMs, only 8 SQOs, 6 clean after anomaly exclusion) — those two could swap or move materially with a few records. paid_social is unmeasurable as given, so any reallocation away from it is contingent on the tracking audit. Pipeline-per-dollar also assumes constant deal size per channel (40k paid_search, 12k linkedin_ads/webinars, 9k organic_search, 8k referral), so it is not validated across deal sizes.
competitor-battlecard-updatePASS 1.00
Plain-English question: Update the Rivally battlecard with sourced pricing and where each side wins.
Prompt — what the model saw
Update the battlecard for the competitor. Sections: one-line positioning, pricing with source and date (newer source wins, note the conflict), where they win, where we win, objections and responses, recent changes, our 12-month win/loss record against them. Cite a snippet id for every factual claim. Rep opinion on a call is not a fact about the competitor. Anything from the old card you cannot re-source gets marked unverified.
cost $0.0232 · 53s · in 3,765 / out 6,169 tok
✓ competitor_named✓ pricing✓ win_loss✓ unverified_marked✓ citations

✓ Correct Answer (ground truth)

Plain-English answer: Rivally at $7 per user per month (the August source wins over the stale $5, conflict noted), a 13-7 win/loss record, and two unverified old claims flagged.
The answer in one line: competitor = Rivally
citation_ids_available25 items — e.g. S01, S02, S03, S04
competitorRivally
pricingvalue_per_user_per_month=7; source_snippet=S17; source_date=2026-08-12
rep_opinion_snippet_ids_not_factsS09, S21
unverified_old_claimsRivally lacks a Slack integration., Rivally was acquired by WorkHuman in 2025.
win_loss_12mowins=13; losses=7
graded checkscompetitor_named · pricing · win_loss · unverified_marked · citations

Model's actual answer

show the model's full answer
# Battlecard: Rivally — updated 2026-10-19

## One-line positioning
Points-based employee recognition for mid-market, expanding into EU enterprise. [S01, S02, S15]

## Pricing
- Current list: $7/user/month (Recognition Starter), annual billing required. Source: pricing page, 2026-08-12 [S17].
- Prior: $5/user/month, annual billing. Source: pricing pages, 2026-01-20 [S03] and 2026-04-01 [S08]. Conflict: S03/S08 show $5, S17 shows $7 — newer source (S17) wins; the $5 figure is outdated, not contradictory.
- Deal quotes: $6.50/user/mo quoted to a 500-seat prospect on annual term, 2026-06-02 [S13] — $1.50 above then-list $5, i.e. +30% (6.50 = 5 × 1.30). Deal mention, not verified list pricing.
- 2026-08-14: prospect reports $7/user/mo list with 15% discount offered for 3-year term → effective ~$5.95/user/mo (7 × 0.85) [S18].
- Note: 2026-06 quote ($6.50) predates and exceeds the then-list $5 — treat as negotiated/seat-specific; no reconciliation available in the data.

## Where they win
- Easy, fast setup; Slack integration works out of the box (mid-market, 2026-02-02) [S04].
- Engaging, points-based recognition feed [S02, S16].
- EU strength: data residency (pitched 2026-02-18 [S05], GA with Dublin office 2026-07-01 [S15]); multi-language praised by EU enterprise reviewer [S12].
- Support response under 4 hours [S22].
- Funding runway for expansion: $40M Series C, Northgate Ventures, 2025-11-04 [S01]; ex-Workday VP EMEA hired 2026-05-09 [S11].

## Where we win
- Analytics depth: G2 reviewer notes limited analytics [S02]; Capterra says dashboards basic vs. enterprise tools [S07]; lost 800-seat deal to Bonusly explicitly citing analytics depth [S25].
- Enterprise admin: no SCIM, manual user management painful [S10]; admin tooling lags peers [S16]; no bulk recognition editing [S24].
- Switching friction as an attack: analytics exports CSV-only made migration off Rivally hard [S20].
- EMEA rewards catalog thinner than US [S14].

## Objections and responses
- "They're cheaper / list is only $5." → $5 is stale; list moved to $7 on 2026-08-12 [S17]. Discounting (15% on 3-year, ~$5.95 effective) exists but is reported via prospect, and the AE's "discounting aggressively" read is rep opinion, not a sourced fact [S18, S21].
- "Their EU data residency beats us." → Real and GA since 2026-07-01 (Dublin office) [S15]; counter with EMEA rewards-catalog gaps [S14].
- "Setup and Slack are easier with them." → Supported by reviews [S04, S02]; counter with admin/SCIM/export gaps at scale [S10, S20, S24].
- "Their UI feels clunky." → Rep opinion only (Bryce Harmon, 2026-04-10), not verified against the product — not a fact; do not use [S09].

## Recent changes
- 2026-08-12: price increase $5 → $7/user/mo on Recognition Starter [S17].
- 2026-09-01: Rivally Pulse exits beta, sold as paid add-on, not bundled [S23]; launched in beta 2026-03-05 [S06].
- 2026-08-19: Microsoft Teams app v2 public preview [S19].
- 2026-07-01: EU data residency GA, Dublin office opened [S15].
- 2026-05-09: ex-Workday VP EMEA hired to lead European expansion [S11].

## Our 12-month win/loss record vs Rivally
- 20 deals on record, Sep 2025 – Aug 2026: 13 wins, 7 losses → 65% win rate (13 ÷ 20 = 0.65).
- By month: 2025-09 1W-1L; 2025-10 2W-0L; 2025-11 1W-1L; 2025-12 1W-1L; 2026-01 2W-0L; 2026-02 2W-0L; 2026-03 1W-1L; 2026-04 0W-2L; 2026-05 0W-1L; 2026-06 1W-0L; 2026-07 1W-0L; 2026-08 1W-0L.
- Downturn: Apr–May 2026 = 0W-3L [deals file], followed by Jun–Aug 2026 = 3W-0L recovery.
- Loss aliases: Deal-7767F5, Deal-5645A5, Deal-C6FFAA, Deal-D263E0, Deal-935746, Deal-9066A6, Deal-72A02F. Win aliases: Deal-A9FD43, Deal-7AA785, Deal-44C524, Deal-0D0CD6, Deal-D5B790, Deal-5C636E, Deal-67BE14, Deal-1B6969, Deal-F03E7B, Deal-072E31, Deal-F65C8F, Deal-E46EAB, Deal-1D2392.
- Note: data covers Sep 2025 – Aug 2026 (12 months as given); Sep 2026 onward has no deals in the file.

## Old-card claims — disposition
- "Rivally lacks a Slack integration" — contradicted: Slack integration worked out of the box [S04]. Removed.
- "Rivally was acquired by WorkHuman in 2025" — unverified: no source in the provided snippets; snippet record shows an independent $40M Series C in Nov 2025 [S01]. Removed from facts; flagged unverified.
- "Pricing starts at $5/user/month (as of 2026-01)" — superseded by S17.
- "Points-based recognition for mid-market" — confirmed [S02].
- "Strong in EU enterprise with multi-language" — confirmed [S12].
nurture-sequence-diagnosisPASS 1.00
Plain-English question: What is broken in each email sequence — rates, overlaps, tracking errors?
Prompt — what the model saw
Review the sequences. Per sequence: sent, open rate, reply rate, meeting rate, weakest step. Identify tracking errors (opened above sent), audience overlap between sequences, and the failure mode for anything under 2 percent reply. One change per weak sequence and which to fix first. Under 350 words.
cost $0.0451 · 76s · in 19,929 / out 8,595 tok
✓ tracking_error✓ overlap_pair✓ weakest_step✓ fix_priority

✓ Correct Answer (ground truth)

Plain-English answer: Per-sequence open, reply, and meeting rates with the weakest step (all three weakest at step 3), an opened-above-sent tracking error, and audience overlap flagged.
The answer in one line: fix_priority_first = Cold Outbound - HR Leaders
all_pairwise_overlaps['Cold Outbound - HR Leaders', 'Cold Outbound - People Ops'], ['Expansion Nurture', 'New Logo Nurture']
audience_overlap_pairCold Outbound - HR Leaders, Cold Outbound - People Ops
failure_modeCold Outbound - HR Leaders=opens healthy (~37%) but replies near zero from step 1 -> body/CTA not landing, not a deliverability or subject-line problem
fix_priority_firstCold Outbound - HR Leaders
per_sequence{"New Logo Nurture": {"sent": 1386, "opened": 490, "replied": 90, "meetings": 27, "open_rate": 0.3535, "reply_rate": 0.0649, "meeting_rate": 0.0195, "weakest_step": 3}, "Expansion Nurture": {"sent": 8
tracking_errorsequence=Expansion Nurture; step=2; sent=300; opened=340; issue=opened exceeds sent
under_2pct_replyCold Outbound - HR Leaders
graded checkstracking_error · overlap_pair · weakest_step · fix_priority

Model's actual answer

show the model's full answer
SEQUENCE FUNNELS (rates vs step sent; totals vs summed sent)

New Logo Nurture: sent 1,386 (500+458+428); open 35.4% (490/1386); reply 6.49% (90/1386); meeting 1.95% (27/1386). Weakest: step 3 (open 28.0%, reply 4.21%, meet 1.40%). Healthy.

Expansion Nurture: sent 875 (300+300+275); open 64.6% (565/875) corrupted; reply 6.74% (59/875); meet 1.37% (12/875). Weakest: step 3 (reply 4.36%, meet 1.09%). Clean open ex-step 2: 39.1% (225/575).

Cold Outbound - HR Leaders: sent 1,785 (600+595+590); open 30.5% (545/1785); reply 0.45% (8/1785); meet 0.00% (0/1785). Weakest: step 3 (reply 0.17%).

Cold Outbound - People Ops: sent 1,163 (400+386+377); open 29.2% (340/1163); reply 2.49% (29/1163); meet 0.52% (6/1163). Weakest: step 3 (reply 1.59%, meet 0.27%).

TRACKING ERRORS
- Expansion Nurture step 2: opened 340 > sent 300 (113.3%) impossible; step 2 sent equals step 1 (300, no decay), consistent with a resend to the full list and/or double-attributed open events. Step 2 open rate unusable.
- All other steps have opened <= sent; no other anomalies.
- audiences.csv rows (299/157/290/217) fall below step-1 sends, so it is a subset; overlap counts are floors.

OVERLAP
- 21 contacts in both Cold Outbound - HR Leaders and Cold Outbound - People Ops: CT-000849, CT-000884, CT-000890, CT-000908, CT-001033, CT-001097, CT-001101, CT-001103, CT-001105, CT-001130, CT-001153, CT-001159, CT-001217, CT-001227, CT-001236, CT-001255, CT-001258, CT-001277, CT-001285, CT-001311, CT-001345.
- 2 contacts in Expansion Nurture and New Logo Nurture: CT-000301, CT-000624.

FAILURE MODE (reply < 2%)
HR Leaders (0.45%, 0 meetings on 1,785 sends): opens are fine (40.0% step 1), so deliverability is not the issue; replies decay 0.83% -> 0.34% -> 0.17% with zero meetings - the subject line earns opens but the offer does not land and no step carries a working CTA. Secondary: the 21 overlapped contacts receive two competing cold emails. Sub-2% steps: all HR Leaders steps; People Ops step 3 (1.59%).

ONE CHANGE PER WEAK SEQUENCE
- HR Leaders: rewrite step 1 around an HR-specific pain with an explicit calendar CTA; pause sends until shipped.
- People Ops: replace step 3 with a new-value step (case study/ROI) instead of a follow-up nudge.
- Expansion: fix step 2 open/send attribution (dedupe the resend) before trusting its funnel.

FIX FIRST: HR Leaders - largest volume (1,785 of 5,209 sends, 34%), zero meetings, and its copy also drags People Ops via the overlap.
weekly-marketing-goals-rollupPASS 1.00
Plain-English question: Are we on pace on the quarter's marketing goals?
Prompt — what the model saw
Produce the weekly marketing goals update for the active quarter: SQMs, SQOs, DS2s, closed-lost MIA rate, same-quarter close count, active pipeline coverage against target. For each: QTD actual, target, delta, pace (on/behind/ahead given days elapsed). One paragraph on what moved this week.
cost $0.0170 · 46s · in 992 / out 5,096 tok
✓ metrics_present✓ deltas✓ coverage_ratio✓ pace✓ paragraph

✓ Correct Answer (ground truth)

Plain-English answer: SQMs ahead of pace, SQOs on pace, DS2s behind (35 short of expected at day 66 of 92), with deltas and same-quarter closes.
The answer in one line: days_in_quarter = 92; days_elapsed = 66; coverage_ratio_active_pipeline = 0.75; elapsed_fraction = 0.7174 (+1 more below)
days_in_quarter92
coverage_ratio_active_pipeline0.75
days_elapsed66
elapsed_fraction0.7174
metrics{"SQMs": {"qtd_actual": 230, "target": 300, "delta": -70, "expected_at_day_elapsed": 215.22, "pace": "ahead"}, "SQOs": {"qtd_actual": 84, "target": 120, "delta": -36, "expected_at_day_elapsed": 86.09,
pace_ruleexpected = target * days_elapsed/days_in_quarter; within +/-5% of expected = on, above = ahead, below = behind; lower_better metrics invert.
this_week_paragraph_required1
graded checksmetrics_present · deltas · coverage_ratio · pace · paragraph

Model's actual answer

show the model's full answer
WEEKLY MARKETING GOALS UPDATE — Q3-2026 (active quarter per quarter_meta.csv)
Quarter: 2026-07-01 to 2026-09-30, 92 days; 66 elapsed. Time elapsed = 66/92 = 71.7%.
Pace rule: expected-at-pace = target x 0.7174; completion vs 71.7%; within +/-2pp = on, above = ahead, below = behind.

SQMs (higher_better)
  QTD actual 230 | target 300 | delta vs target -70 (76.7% of target)
  Expected at pace 300 x 0.7174 = 215.2 | delta vs pace +14.8
  Pace: AHEAD (76.7% vs 71.7%)

SQOs (higher_better)
  QTD actual 84 | target 120 | delta vs target -36 (70.0% of target)
  Expected at pace 120 x 0.7174 = 86.1 | delta vs pace -2.1
  Pace: ON (70.0% vs 71.7%, at the edge of the band)

DS2s (higher_better)
  QTD actual 40 | target 75 | delta vs target -35 (53.3% of target)
  Expected at pace 75 x 0.7174 = 53.8 | delta vs pace -13.8
  Pace: BEHIND (53.3% vs 71.7%)

Closed-lost MIA rate (lower_better)
  QTD actual = 5/25 = 20.0% | target 10.0% | delta +10.0pp (2.0x target, unfavorable)
  Pace: BEHIND. Rate is a ratio of closed-lost to date, not an accumulating flow, so day-proration does not apply; it stands at twice the target.

Same-quarter closes (higher_better)
  QTD actual 10 | target 20 | delta vs target -10 (50.0% of target)
  Expected at pace 20 x 0.7174 = 14.3 | delta vs pace -4.3
  Pace: BEHIND (50.0% vs 71.7%)

Active pipeline (higher_better)
  QTD actual 3,000,000 | target 4,000,000 | coverage = 3,000,000/4,000,000 = 75.0% | delta vs target -1,000,000
  Pace: AHEAD nominally (75.0% vs 71.7%, +3.3pp), with the caveat that pipeline is a point-in-time stock, so linear time-proration is weak; the operative read is 75% coverage, $1.0M short of target.

WHAT MOVED THIS WEEK
The provided files contain only quarter-to-date cumulative values (marketing_qtd.csv) plus targets and quarter metadata — there is no prior-week or week-over-week snapshot, so week-over-week movement cannot be computed from the given data and none is asserted here. On the QTD standing at day 66: SQMs are the only flow metric comfortably ahead (230 vs 215 expected), SQOs sit right at the pace line (84 vs 86 expected), and the quarter's shortfall is concentrated downstream of the SQO step — DS2s at 53.3% of target and same-quarter closes at 50.0%, both well behind the 71.7% time mark — while closed-lost MIA rate is at 20.0% against a 10.0% target (5 of 25 closed-lost), twice the goal. Pipeline coverage stands at 75% of the $4.0M target ($3.0M active, $1.0M short). To report actual weekly movement, a previous week's QTD snapshot would be needed.
communication 5 tests
ceo-slack-compressionPASS 1.00
Plain-English question: Compress the forecast analysis into one Slack message the CEO can read in ten seconds.
Prompt — what the model saw
Compress the forecast analysis below into a single Slack message for the CEO. Rules: at most three sentences; under 80 words; lead with the decision or the so-what; keep every distinct number that matters and the one caveat that changes the conclusion; cut any sentence that does not carry a claim, a number, or a directive. No em dashes, no throat-clearing phrases.
cost $0.0138 · 50s · in 1,360 / out 3,935 tok
✓ max_3_sentences✓ under_80_words✓ weighted_present✓ caveat_present✓ no_em_dash

✓ Correct Answer (ground truth)

Plain-English answer: Three sentences, under 80 words, decision first, every distinct number kept plus the one caveat — all checked by deterministic text rules.
The answer in one line: weighted_number = 115976; must_keep = data-quality caveat — deleting the bad news fails
weighted_number115,976
limitmax 3 sentences, max 80 words, no em-dashes
must_keepdata-quality caveat — deleting the bad news fails
graded checksmax_3_sentences · under_80_words · weighted_present · caveat_present · no_em_dash

Model's actual answer

show the model's full answer
Q3 forecast is $115,976.75, but treat it as soft: all 7 in-quarter COMMIT deals have empty why-buys, so the committed base has no documented reason to buy. Weighting: $44,729 COMMIT at 100% plus 35% of $203,565 BEST_CASE (24 deals) = $71,247.75. Excluded: 32 deals worth $227,575 close after 2026-09-30, including $13,770 COMMIT Deal-D348E1 dated 2026-10-15.
stalled-deal-followup-emailPASS 1.00
Plain-English question: Write a follow-up email for this stalled deal — under 120 words, one prior touch, one ask.
Prompt — what the model saw
Write a follow-up email for this stalled deal. Under 120 words. Reference one specific prior touch from the engagement history, make one clear ask, and match the tone of the prior thread. No invented details.
cost $0.0126 · 36s · in 1,127 / out 3,590 tok
✓ word_cap✓ references_touch✓ single_ask

✓ Correct Answer (ground truth)

Plain-English answer: An email that references the July 28 milestone-awards demo, makes one clear ask, and invents nothing — Deal-0D2F7A, 31 days since last engagement.
The answer in one line: amount = 5100; word_cap = 120; days_since_last_engagement = 31; deal_alias = Deal-0D2F7A (+4 more below)
amount5,100
days_since_last_engagement31
word_cap120
allowed_facts14 items — e.g. Deal-0D2F7A, $5,100, DS3, Alex Franklin
allowed_newENG-00001, ENG-00002, ENG-00003, ENG-00004, ENG-00005
deal_aliasDeal-0D2F7A
must_referenceengagement_id=ENG-00002; date=2026-07-28; type=meeting; why=most substantive prior touch; concrete prospect reactions to milestone awards and points catalog
ownerAlex Franklin
single_cta1
stageDS3
tonewarm, concise, professional (matches prior thread)
graded checksword_cap · references_touch · single_ask

Model's actual answer

show the model's full answer
Subject: Following up on the July 28 demo pricing

Hi [Name],

I'm following up on my Aug 5 email that recapped the July 28 demo and included pricing for 150 seats.

Since it's been a couple of months, I wanted to check whether the proposal still fits your plans, and to offer a quick walk-through of any pricing questions.

Could you let me know if you have 15 minutes in the next two weeks to review the pricing together?

Best,
Alex Franklin

Reference: ENG-00001 (Deal-0D2F7A), last engagement 2026-08-05; gap to today (2026-10-19) is ~2.5 months. Word count: 8 + 2 + 20 + 28 + 20 + 3 = 81 words.
funnel-digest-paragraphPASS 1.00
Plain-English question: Write the weekly GTM digest — one paragraph per team, every number traceable.
Prompt — what the model saw
Write the weekly GTM digest: one paragraph per team (Marketing, Sales, CS, PLG) in Bonusly brand voice. Every number must trace to the extracts. No customer contact names, no employee performance commentary, no compensation. Four paragraphs, under 400 words total.
cost $0.0240 · 38s · in 11,034 / out 3,472 tok
✓ four_paragraphs✓ word_cap✓ teams

✓ Correct Answer (ground truth)

Plain-English answer: Four paragraphs (Marketing, Sales, CS, PLG) using only the allowed numbers, no contact names, no performance commentary, under 400 words.
The answer in one line: word_cap = 400
word_cap400
allowed_numbers16 items — e.g. 2, 3, 9, 08
bannedcustomer contact names, employee performance commentary, compensation
required_paragraphsMarketing, Sales, CS, PLG
graded checksfour_paragraphs · word_cap · teams

Model's actual answer

show the model's full answer
Weekly GTM digest, week of 2026-08-31

Marketing. Demand generation came up close, not home. The team delivered 46 SQMs against a target of 52, a gap of 6 (52 − 46 = 6) and 88% attainment (46 ÷ 52 ≈ 0.88). Webinars did the heavy lifting with 18 of those 46 SQMs, close to 4 in 10 (18 ÷ 46 ≈ 0.39). Paid search cost per SQM held at $150, a steady read on efficiency.

Sales. The pipeline engine kept turning. Sales converted 14 SQOs and set 9 DS2 meetings this week. New pipeline created was $310,000, and the same-quarter close count for the week was 3.

CS. Customer Success saved 2 renewals this week and moved Team NPS to 61. The watch item going into next week: 3 open red-flag accounts.

PLG. The self-serve funnel kept growing. PLG added 412 new signups with activation at 31 percent, roughly 128 activated accounts (412 × 0.31 ≈ 128). 38 companies reached the aha moment of 10 recognition gives, about 9 percent of new signups (38 ÷ 412 ≈ 0.09).

Note: the extracts give no prior-week baselines (NPS, pipeline, or signups), so week-over-week changes cannot be computed; all figures above trace to the week_extracts.md values as given.
partner-digest-sectionPASS 1.00
Plain-English question: Write the partner digest so every partner in the roster appears exactly once.
Prompt — what the model saw
Write the partner digest for the period. Every partner in the roster appears once: active plays get status with numbers, quiet partners get one line. Partner-sourced pipeline count and amount must match the deal data. Nothing omitted, nothing added.
cost $0.0073 · 18s · in 1,172 / out 1,800 tok
✓ all_partners✓ pipeline_amount✓ quiet_one_line

✓ Correct Answer (ground truth)

Plain-English answer: Four active partners with status and numbers, four quiet partners with one line each, and partner-sourced pipeline that reconciles to the deal data.
The answer in one line: rule = every roster partner named exactly once; quiet partners get one line; no partners outside roster.
active_no_dealsWorkWell Group
partner_sourced_pipelinecount=5; amount=550000
per_partner_pipeline{"Apex Rewards Co": {"count": 2, "amount": 275000}, "HRCloud Partners": {"count": 1, "amount": 140000}, "CultureBridge": {"count": 2, "amount": 135000}}
quiet_partnersRecogniq, KudosWave, PeopleFirst Advisors, TotalPerk
rosterpartner_name=Apex Rewards Co; status=active | partner_name=HRCloud Partners; status=active | partner_name=CultureBridge; status=active | partner_name=WorkWell Group; status=active | partner_name=Recogniq; status=quiet | partner_name=KudosWave; status=quiet … (+2 more)
ruleevery roster partner named exactly once; quiet partners get one line; no partners outside roster.
graded checksall_partners · pipeline_amount · quiet_one_line

Model's actual answer

show the model's full answer
Partner digest — period 2026-08-24 to 2026-09-04

Partner-sourced pipeline (utm_source = Partner): 5 deals, $550,000 total.

Active partners

Apex Rewards Co — active play. 2 sourced deals: Deal-DDAAF2 ($180,000) + Deal-2CF33E ($95,000) = $275,000. Both logged with UTM Source = Partner and both in DS1 per Slack (08-25). Co-webinar locked for 09-15.

HRCloud Partners — active play. 1 sourced deal: Deal-F1CDA5 ($140,000). Security review closed via the integration referral; opp moved to DS2 per Slack (08-27).

CultureBridge — active play. 2 sourced deals: Deal-096E1D ($60,000) + Deal-067213 ($75,000) = $135,000. Both produced by the lunch-and-learn, both early stage per Slack (08-29).

WorkWell Group — active play, no sourced deals this period ($0, 0 deals). Wants to restart the joint playbook in Q4; planning call booked for 09-09 per Slack (09-02).

Quiet partners

Recogniq — quiet; no Slack activity and no deals in the period.
KudosWave — quiet; no Slack activity and no deals in the period.
PeopleFirst Advisors — quiet; no Slack activity and no deals in the period.
TotalPerk — quiet; no Slack activity and no deals in the period.

Totals and reconciliation

- Deal counts: Apex 2 + HRCloud 1 + CultureBridge 2 + WorkWell 0 + quiet partners 0 = 5 deals.
- Amounts: $275,000 (Apex) + $140,000 (HRCloud) + $135,000 (CultureBridge) + $0 (WorkWell) = $550,000.
- Slack-reported sourced-op counts (Apex 2, HRCloud 1, CultureBridge 2, WorkWell 0) match the deal file exactly for every partner.
rfp-answer-from-policyPASS 1.00
Plain-English question: Answer the 10 RFP items from the policy excerpts only — no guessing.
Prompt — what the model saw
Answer the 10 questionnaire items from the policy excerpts only. Cite the excerpt id for every answer. Where the excerpts do not answer the question, mark it unanswerable and say what document would; do not guess. Tone: vendor response, direct.
cost $0.0046 · 12s · in 1,149 / out 915 tok
✓ answered_count✓ unanswerable_flagged✓ citations✓ no_guess

✓ Correct Answer (ground truth)

Plain-English answer: Seven answered with excerpt citations (Q1-Q7), three flagged unanswerable with the document that would answer them (Q8-Q10).
The answer in one line: answered_count = 7; rule = unanswerable items must be flagged and the document that would answer them named; do not guess.
answered_count7
answer_map{"Q1": ["P1"], "Q2": ["P1"], "Q3": ["P2"], "Q4": ["P3"], "Q5": ["P4"], "Q6": ["P5"], "Q7": ["P6"]}
ruleunanswerable items must be flagged and the document that would answer them named; do not guess.
unanswerableQ8, Q9, Q10
valid_citation_idsP1, P2, P3, P4, P5, P6, P7, P8
graded checksanswered_count · unanswerable_flagged · citations · no_guess

Model's actual answer

show the model's full answer
Q1 — Yes. Customer data is encrypted at rest using AES-256. [P1]

Q2 — Yes. Data in transit is encrypted with TLS 1.2 or higher. [P1]

Q3 — Yes. Bonusly maintains a SOC 2 Type II report, renewed annually. The report is available to prospects under NDA. [P2]

Q4 — Customer data is retained for the duration of the contract plus 90 days, after which it is deleted from production systems. [P3]

Q5 — The current subprocessor list is maintained at trust.bonusly.com. Customers receive 30 days notice before a new subprocessor is added. [P4]

Q6 — Yes. SAML 2.0 single sign-on is available on the Enterprise plan (SCIM user provisioning is also available on that plan). [P5]

Q7 — Production databases are backed up daily. Backups are encrypted and retained for 35 days. [P6]

Q8 — Unanswerable from the provided excerpts. The excerpts address SOC 2 only [P2] and make no reference to ISO 27001. The ISO 27001 certificate or the trust portal's compliance certification page would answer this.

Q9 — Unanswerable from the provided excerpts. No uptime or availability commitment appears in them. The executed Master Services Agreement / Service Level Agreement (or the trust portal SLA page) would state the contractual uptime SLA percentage.

Q10 — Unanswerable from the provided excerpts. Nothing in them addresses HIPAA or Business Associate Agreements. The Bonusly BAA template / legal and compliance team (or the trust portal's compliance coverage page) would confirm whether a HIPAA BAA is offered and under what terms.

Note on arithmetic: no calculations were required — all answers are direct citations. Excerpt coverage: Q1–Q7 answered from [P1]–[P7]; Q8–Q10 have no supporting excerpt.
ops-maintenance 5 tests
skill-file-reconciliationPASS 1.00
Plain-English question: Audit a skill set against its manifest and flag every conflict with a severity and an action.
Prompt — what the model saw
Reconcile the skill set below against its manifest. Find and report (with severity CRITICAL/WARNING/INFO and an action type TRIM_DESC, UPDATE_BODY, DELETE_SKILL, MERGE, or REVIEW): (1) any two skills whose ALWAYS-trigger phrases overlap/duplicate, (2) any circular delegation chain between skills (name it), (3) any delegation target that does not exist (dangling, name it), (4) any version conflict and which skill should survive, (5) how many manifest descriptions exceed 1,024 characters, (6) any hardcoded page ids, dates, or person names in skill bodies, (7) manifest drift in BOTH directions (files with no manifest row, and manifest rows with no file). Make one proposal per finding; rewrite nothing.
cost $0.1469 · 220s · in 91,830 / out 27,056 tok
✓ duplicate_trigger_found✓ circular_chain✓ dangling_target✓ oversized_desc_count✓ hardcoded_values✓ manifest_drift✓ severity_and_action

✓ Correct Answer (ground truth)

Plain-English answer: A severity-ranked list of TRIM_DESC / UPDATE_BODY / DELETE_SKILL / MERGE / REVIEW actions covering every rule violation found.
The answer in one line: files_audited = 14; rank = CRITICAL / WARNING / INFO with one fix action per issue
files_audited14
findingsduplicate trigger, circular delegation chain, dangling file reference, oversized description, hardcoded values, manifest drift
rankCRITICAL / WARNING / INFO with one fix action per issue
graded checksduplicate_trigger_found · circular_chain · dangling_target · oversized_desc_count · hardcoded_values · manifest_drift · severity_and_action

Model's actual answer

show the model's full answer
Reconciliation report. All findings based solely on the provided manifest and skill bodies. Arithmetic shown where relevant. No rewrites made.

FINDINGS

1. CRITICAL / MERGE — Duplicate ALWAYS-trigger phrases: comms-drafter vs email-drafter.
Both claim the same trigger phrases verbatim: "write me an email," "draft a follow-up," "what should I say," "bump email," "contract nudge," and both fire on "pastes an existing message/email and asks for feedback." Both target the same job (customer-facing email drafting). Routing is ambiguous and outputs would duplicate.
Proposal: merge the two into one drafting skill; comms-drafter should be the surviving surface (superset: email + support/Intercom + partner comms + brand-first step), with email-drafter's two unique sections (Gmail signature retrieval, user identity/scoping via get_user_details) folded in before retirement.

2. WARNING / REVIEW — Near-duplicate trigger phrases: weekly-pipeline-report vs pipeline-intelligence-report.
Overlaps: weekly's "update the pipeline" ≈ PIR's "pipeline update"; weekly's "what does pipeline look like" ≈ PIR's "what's the pipeline look like"; both claim "the pipeline report" variants ("generate the pipeline report" / "run the pipeline report"). Both are pipeline report generators with different outputs (HTML weekly vs 10-tab scored HTML).
Proposal: disambiguate triggers — reserve "pipeline update"/"what's the pipeline look like" wording to one skill and prefix the other with its cadence ("weekly") — then re-review routing.

3. WARNING / REVIEW — Trigger overlap: deal-strategy-coach vs next-to-close.
deal-strategy-coach triggers on manager asks including "which deals are likely to close"; next-to-close ALWAYS-triggers on "which deals are most likely to close." Same phrasing family, two skills.
Proposal: scope the deal-strategy-coach trigger to coaching context only and leave shortlist phrasing to next-to-close.

4. INFO / REVIEW — Overlapping always-mandatory mandates over the same SignalForge outputs.
Four skills each claim unconditional invocation over the same output stream: model-selection ("ALWAYS run... at the start of every task"), analysis-validator ("Runs after every SignalForge quantitative analysis... Never skip"), signalforge-claim-compressor ("Final style pass... ALWAYS trigger for: SignalForge reports..."), signalforge-feedback ("ALWAYS trigger... absolute final step... after analysis-validator and after signalforge-claim-compressor"). Sequencing is held only by prose convention in each body.
Proposal: move ordering ownership to a single orchestrator row (see finding 6) and reduce each body's trigger to its own phase.

5. WARNING / UPDATE_BODY — Circular delegation chain: email-drafter → deal-strategy-coach → email-drafter.
email-drafter's lane marker sends strategy questions to deal-strategy-coach; deal-strategy-coach's manager-email section sends drafting back to email-drafter ("use the email-drafter skill"). A request that mixes manager-email drafting with strategy can ping-pong between the two lane markers. (comms-drafter → deal-strategy-coach is one-way and not part of the cycle. next-to-close → pipeline-intelligence-report → closed-lost-analysis is one-way: closed-lost-analysis only notes being "called from" PIR, it does not delegate back. No other cycles found.)
Proposal: give deal-strategy-coach ownership of strategy output and email-drafter exclusive ownership of final drafting, removing the reciprocal pointer that closes the loop.

6. WARNING / REVIEW — Dangling delegation targets (not in manifest, not provided):
- bonusly-brand — referenced by comms-drafter, email-drafter, sales-forecast, signalforge-claim-compressor
- prospect-research-multithreading — referenced by comms-drafter, deal-strategy-coach, email-drafter
- skill-orchestrator — referenced by analysis-validator (self-improvement cascade), signalforge-feedback (registration checklist)
- signalforge-reports (as /mnt/skills/organization/signalforge-reports/SKILL.md) — referenced by pipeline-intelligence-report, weekly-pipeline-report
- bonusly-data-questions, bonusly-product-questions, bonusly-business-reporting-questions, bonusly-rewards-questions, bonusly-ppp-questions, bonusly-feature-flag-questions, bonusly-deal-desk-questions, bonusly-datadog-questions — delegation table in analysis-validator §12.4
- deal-strategy-coach and pipeline-intelligence-report are NOT dangling (in manifest); comms-drafter's "use deal-strategy-coach" is valid.
Proposal: either add manifest rows for these external skills or annotate each reference as out-of-manifest/external so dangling pointers are distinguishable from unregistered ones.

7. WARNING / UPDATE_BODY — Version conflict inside analysis-validator: v3.6 vs v3.2.
Header/changelog say Version 3.6 (May 9, 2026); the Section 7 validation-trail template still prints "Validator: analysis-validator v3.2." pipeline-intelligence-report's footer also stamps "Analysis Validator v3.6." v3.6 is the version that should survive; the trail template line is stale.
Proposal: update the trail template to render the current version variable.

8. WARNING / REVIEW — Roster conflict: analysis-validator vs pipeline-intelligence-report.
analysis-validator §12.3 lists "Core 6 AEs" (Bryce Harmon, Hugo Lindqvist, Dana Mercer, Alex Franklin, Cole Ingram, Gavin Porter) and states any "full AE team"/"Core 6" filter must include all six IDs. pipeline-intelligence-report Phase 1 "AE owner IDs (verified May 2026)" lists only five — Hugo Lindqvist (77260721) is missing. The analysis-validator roster (Updated May 4, 2026) should survive as the canonical roster; PIR's embedded list is the stale copy.
Proposal: PIR should resolve owners dynamically (as stale-pipeline-report Phase 2 already does) instead of embedding a roster.

9. WARNING / UPDATE_BODY — sales-forecast internal version conflict: v1.1 claims quarter-agnostic but Tab 6 is hardcoded "Q2 Narrative."
Changelog v1.1 (April 27, 2026): "Quarter-agnostic (Q2 → current quarter throughout)." The tab table still names tab 6 "Q2 Narrative." The quarter-agnostic intent (v1.1) should survive.
Proposal: rename tab 6 to "Q[N] Narrative."

10. INFO / TRIM_DESC — Descriptions exceeding 1,024 characters: zero (0).
Arithmetic: declared description_chars across 14 rows: 656, 897, 996, 792, 965, 676, 945, 1004, 1006, 962, 1006, 708, 762, 656. Count > 1,024 = 0. Maximum = 1,006 (pipeline-intelligence-report and signalforge-claim-compressor), 1,006 ≤ 1,024. Spot-check against body text: next-to-close measured 945 = declared 945; deal-strategy-coach 792 = 792; email-drafter 965 (±1 by transcription) — counts consistent.
Proposal: no trim required; flag the two 1,006-char rows as 18 characters from the cap before any future edit.

11. WARNING / UPDATE_BODY — Hardcoded ids in skill bodies.
- partner-digest: Confluence Cloud ID 73fe98de-a4a3-4869-9f8a-bb1eeed4cf7f, Space ID 1958248479, folder ID 2286616609, page URLs 2286321666, 2265382925, 2236940297, 2237825028, 2239365136, 2238283777; Slack user <@U03QLMBL7AR>.
- signalforge-feedback: page IDs 2295136266, 2234417154, 2247295002; spaceId 2232811524; Cloud ID 73fe98de-…
- sales-forecast: Space ID 2232811524, Cloud ID, parent page 2232582148.
- deal-strategy-coach: Confluence page 2257879045 in AE Excellence Playbook URL.
- weekly-pipeline-report: Google Sheets IDs 1CLZeOsElVDF_LF0ZG_t2nfwvhnZ6bpwqM_nX3WEYzcw and 1ENuaEcCuLjdKhMvp8FK3Ys1ek5Aw9ZuOZhsHJJFoB_k.
- stale-pipeline-report: Slack channel C0561C1JCPJ, HubSpot org 1973303, owner ID 55483190.
- pipeline-intelligence-report, next-to-close, stale-pipeline-report, weekly-pipeline-report, analysis-validator: HubSpot org ID 1973303 and stage IDs 150582536–150582539, 1175632767 embedded in multiple places (stage map duplicated in 5 bodies).
Proposal: externalize ids into per-skill reference/config files (or a shared constants reference) so single values live in one place.

12. WARNING / UPDATE_BODY — Hardcoded dates in skill bodies.
- model-selection: last_checked 2026-05-19. Its own staleness rule says re-check if today is >14 days past that date. Today 2026-10-19 is 153 days past (31+30+31+31+30), so the registry is self-declared stale right now.
- weekly-pipeline-report: "Q2 (April 1 – June 30, 2026; total ≈ 64–65)" business-day computation and static Q1 2026 actuals ($365,152 vs $475,000 plan, 77%; $2,490,532 vs $3,288,000 forecast, 76%) — quarter-locked despite "always read live" instructions.
- analysis-validator: dated constants (May 4 2026 CALL_SPOTLIGHT_BRIEF removal, March 28 2023 stale DEALS table, May 2026 anchors "3,000–3,500 paying customers / 440,000–470,000 users," "as of May 2026" ranges).
- closed-lost-analysis: "30-deal AI-field sample from May 2026: 10 of 10," "demo on 4/13… deal closed in May," "rep vacation May 4–12," "May 2026" field confirmations.
- partner-digest: example dates May 19 / June 2 / May 16, 2026 and the May 16, 2026 canonical reference issue.
Proposal: replace fixed dates with relative windows or move them into a dated-reference file, keeping only the self-updating rule (model-selection) machine-checkable.

13. WARNING / UPDATE_BODY — Hardcoded person names in skill bodies.
- analysis-validator §12.3: full named rosters with owner IDs (Bryce Harmon, Hugo Lindqvist, Dana Mercer, Alex Franklin, Cole Ingram, Gavin Porter, Colleen Perry, Ellie Barton, Ashley Reyer, Megan Franz, Elena Sinclair, Youssef Elkhateeb, Amanda Czenkus, Alaina Loori, Shealagh Coughlin, Ben Castelli, Amani Phipps, John Thomas, Yasmin Wahid); §10/G1-K escalation names "Manish or Amani"; §12.2/§13 example names "Dana Mercer," "Gavin Porter."
- pipeline-intelligence-report Phase 1: five AE names with owner IDs.
- partner-digest: "Owner: Amani Phipps," Slack identity "Amani's threads"/from:<@U03QLMBL7AR>, contacts "Kelli, Jen Lee," "Hani, Bryce," "Sara."
- weekly-pipeline-report: title "Ben Lavin · Demand Generation"; "presented in chat for Ben's review."
- deal-strategy-coach: ".edu … routed to Farid," "India (routed to Perseus)."
- sales-forecast changelog: "Elena → Alaina (VP Sales)"; signalforge-feedback examples: "Gavin Porter Rep Diagnostic," "Lowe's Conversation Analysis."
- closed-lost-analysis: named accounts/competitors as illustrative examples (Softheon, Estee Lauder, LIFTOFF, Nestlé, Ozinga, MinIO, Aurora Innovation, GCash, Ethos Cannabis, StickerYou); signalforge-claim-compressor examples (Felix Construction, Panopto, Schneider Downs).
Proposal: move rosters/named contacts to a reference file resolved at runtime (stale-pipeline-report Phase 2 already demonstrates the pattern: "Never hardcode rep names or owner IDs"), and replace illustrative customer names with anonymized placeholders.

14. INFO / no action — Manifest drift in both directions: zero.
Arithmetic: manifest rows = 14; .SKILL.md files provided = 14. Set difference (files with no row) = ∅; set difference (rows with no file) = ∅; every declared_name equals its file stem. Spot-check of declared description_chars against actual body text matched (see finding 10).
Proposal: none — drift-free as provided.

15. INFO / REVIEW — Missing linked references prevent verification.
sales-forecast depends on references/data-sources.md, references/report-structure.md, references/cadence.md, references/report-template.html, references/TEMPLATE_README.md; weekly-pipeline-report depends on references/report-spec.md, references/queries.md; stale-pipeline-report invokes /mnt/skills/public/xlsx/scripts/recalc.py. None of these are included in the provided bundle, so their content (scoring formulas, query specs, template rules) cannot be reconciled.
Proposal: include references/ files in the manifest scope (or mark them out-of-scope) before the next reconciliation.

SUMMARY OF SEVERITIES: 1 CRITICAL (finding 1), 9 WARNING (findings 2–9, 11–13), 5 INFO (findings 4, 10, 14, 15 plus the headroom note in 10). Actions used: MERGE (1), REVIEW (5), UPDATE_BODY (5), TRIM_DESC-flag only (0 actual trims needed), DELETE_SKILL (0 — the comms-drafter/email-drafter case is handled as MERGE per finding 1; if merge is rejected, DELETE email-drafter after porting its unique sections). Data not provided in the bundle (references/ files, model registry currentness, live HubSpot rosters) is flagged rather than assumed.
runbook-from-slack-threadPASS 1.00
Plain-English question: Turn the resolved incident thread into a runbook.
Prompt — what the model saw
Turn the resolved incident thread into a runbook: ordered steps with the exact command or action taken, who ran it, how success was verified, and rollback for each step that changed state. Every step must trace to a message in the thread. Add nothing the thread does not contain; where the thread is unclear, mark the step as needing confirmation.
cost $0.0215 · 46s · in 3,757 / out 5,659 tok
✓ ordered_steps✓ rollback✓ needs_confirmation✓ sourced

✓ Correct Answer (ground truth)

Plain-English answer: Ordered steps from the thread, each with the exact command, who ran it, how success was verified, and rollback where state changed.
The answer in one line: rule = no steps beyond what the thread contains; every step traces to a message id.
ambiguous_steporder=4; msg=M05; mark=needs-confirmation
ruleno steps beyond what the thread contains; every step traces to a message id.
stepsorder=1; msg=M01; actor=Bryce Harmon; action=acknowledge PagerDuty alert, take IC; state_change=False; verified_by=M01 | order=2; msg=M02; actor=Farid Osman; action=check queue depth; command=bundle exec rake sidekiq:queue_depth; state_change=False; verified_by=M02 (48,213 pending) | order=3; msg=M04; actor=Farid Osman; action=pause auto-enqueue via feature flag; command=bin/rails runner 'FeatureFlag.disable(:auto_recognition_enqueue)'; state_change=True; rollback=bin/rails runner 'FeatureFlag.enable(:auto_recognition_enqueue)'; verified_by=M07 (queue depth falling) | order=4; msg=M05; actor=Elena Sinclair; action=cleared the dead set; state_change=True; needs_confirmation=thread gives no command or method for clearing the dead set | order=5; msg=M06; actor=Bryce Harmon; action=scale reward-worker to 6 replicas; command=kubectl scale deployment/reward-worker --replicas=6; state_change=True; rollback=kubectl scale deployment/reward-worker --replicas=3; verified_by=M07/M08 | order=6; msg=M08; actor=Cole Ingram; action=verify queue drained and error rate baseline; command=bundle exec rake sidekiq:queue_depth; state_change=False; verified_by=M08 (returns 0) … (+2 more)
graded checksordered_steps · rollback · needs_confirmation · sourced

Model's actual answer

show the model's full answer
Runbook — reward-worker queue backlog, incident-2026-08-29-reward-queue (resolved 2026-08-29)
All steps trace to messages M01–M10 of the supplied thread. State changes flagged; every gap marked "needs confirmation."

1. Acknowledge alert, take IC — Bryce Harmon [M01, 14:02:10Z]
   Action: PagerDuty alert fired for reward-worker queue depth > 10k; Bryce Harmon acknowledged and took IC.
   Verification: None documented (self-acknowledgment only).
   Rollback: Not a system state change.

2. Check queue depth — Farid Osman [M02, 14:04:33Z]
   Command: bundle exec rake sidekiq:queue_depth
   Result: reward queue at 48,213 pending jobs; normal is under 500.
   Verification: Direct — command output in thread.
   Rollback: None (read-only).

3. Inspect dead set — Farid Osman [M03, 14:06:02Z]
   Action: Reported dead set has 112 jobs, all Redis::TimeoutError from around 13:58.
   Verification: Reported result; the inspection command is not in the thread — needs confirmation.
   Rollback: None (read-only).

4. Pause enqueue — Farid Osman [M04, 14:08:45Z] — STATE CHANGE
   Command: bin/rails runner 'FeatureFlag.disable(:auto_recognition_enqueue)'
   Verification: No direct verification that the flag flipped is recorded — needs confirmation.
   Rollback (documented in M04): bin/rails runner 'FeatureFlag.enable(:auto_recognition_enqueue)'

5. Clear dead set — Elena Sinclair [M05, 14:15:20Z] — STATE CHANGE (destructive)
   Action: Elena Sinclair states she "cleared out the dead set" while "in the console." Exact command not documented — needs confirmation.
   Verification: No independent verification documented — needs confirmation.
   Rollback: Not documented — needs confirmation. Do not treat as an approved repeatable remediation without a documented command, verification, and rollback.

6. Scale workers up — Bryce Harmon [M06, 14:21:07Z] — STATE CHANGE
   Command: kubectl scale deployment/reward-worker --replicas=6 (was 3)
   Verification: No direct replica-count verification recorded. Later queue observations (M07, M08) are not isolated proof of this change — needs confirmation.
   Rollback (documented in M06): kubectl scale deployment/reward-worker --replicas=3

7. Monitor queue drain — Farid Osman [M07, 14:33:41Z]
   Result: Queue depth reported at 9,400 and falling ~1,200/min.
   Verification: Measurement command not documented — needs confirmation. This is a reported observation, not verification of M04 or M06.
   Rollback: None (read-only).

8. Verify drain complete — Cole Ingram [M08, 14:47:55Z]
   Command: bundle exec rake sidekiq:queue_depth — returns 0.
   Additional: Datadog error rate reported back to baseline (check method not documented in thread).
   Verification: Direct for queue depth; Datadog portion is reported observation.
   Rollback: None (read-only).

9. Re-enable enqueue — Bryce Harmon [M09, 14:49:10Z] — STATE CHANGE
   Command: bin/rails runner 'FeatureFlag.enable(:auto_recognition_enqueue)'
   Verification: Reported 40 new jobs processed cleanly in the next 3 minutes (verification method not documented in thread).
   Rollback: Not documented — needs confirmation (do not infer disabling the flag as the approved rollback).

10. Scale workers back down, close incident — Bryce Harmon [M10, 14:55:00Z] — STATE CHANGE
    Command: kubectl scale deployment/reward-worker --replicas=3
    Verification: Reported "queue stable at 0"; no command quoted in this message, and no replica-count check documented — needs confirmation.
    Rollback: Not documented — needs confirmation (do not infer scaling to 6 as the approved rollback).
    Outcome per thread: incident resolved.

Open gaps carried from the thread (all marked needs confirmation above): inspection command in M03; direct flag verification for M04; command, verification, and rollback for the dead-set clear in M05; replica-count verification for M06 and M10; measurement command in M07; verification method behind the "40 new jobs" check in M09; rollback for M09 and M10.
datadog-incident-timelinePASS 1.00
Plain-English question: Sequence the failure from the logs — what broke first, and what cascaded after it?
Prompt — what the model saw
From the log slice, sequence the failure: first error with timestamp, the cascade in order, the service and job involved, and the Datadog query that would confirm the first error. State what the logs do not show.
cost $0.0427 · 80s · in 8,131 / out 11,820 tok
✓ first_error_ts✓ service_named✓ cascade_ordered✓ query✓ gaps_stated

✓ Correct Answer (ground truth)

Plain-English answer: A reward-service Redis timeout at 14:01:12Z cascades to sidekiq at 14:01:40 and api-gateway at 14:03:05, with the Datadog query that confirms the first error.
The answer in one line: job_class_involved = RewardGiveJob
cascade_orderservice=reward-service; first_error_at=2026-09-03T14:01:12Z | service=sidekiq; first_error_at=2026-09-03T14:01:40Z | service=api-gateway; first_error_at=2026-09-03T14:03:05Z | service=web-app; first_error_at=2026-09-03T14:03:30Z
clean_servicespostgres
confirming_queryservice:reward-service status:error "@message:Redis::TimeoutError"
first_errortimestamp=2026-09-03T14:01:12Z; service=reward-service; message=Redis::TimeoutError: Connection to Redis at redis-primary:6379 timed out after 5s
job_class_involvedRewardGiveJob
logs_do_not_showroot cause of the Redis timeout, redis-server's own logs, infrastructure metrics (CPU/memory/network), any deploy or config change preceding the incident, who applied mitigation or when (no actor fields in logs)
graded checksfirst_error_ts · service_named · cascade_ordered · query · gaps_stated

Model's actual answer

show the model's full answer
Failure sequence, 2026-09-03, all times UTC. Only the two provided files were used.

FIRST ERROR
2026-09-03T14:01:12Z — reward-service, ERROR: "Redis::TimeoutError: Connection to Redis at redis-primary:6379 timed out after 5s". It is the first ERROR line in the slice; the last preceding log is an INFO at 13:59:30 ("job enqueued", reward-service), a 102s gap (13:59:30 -> 14:01:12).

CASCADE IN ORDER
1. 14:01:12 — reward-service: Redis connection to redis-primary:6379 times out after 5s (implies attempt start ~14:01:07; that start is not logged).
2. 14:01:20, 14:01:30, 14:01:40 — reward-service: "retry exhausted for RewardGiveJob" x3.
3. 14:01:40 — sidekiq: "RewardGiveJob failed: Redis::TimeoutError; retrying in 60s". sidekiq_jobs.csv logs the first batch J-00001..J-00006 at 14:01:46-14:01:57 (6 RewardGiveJob failures).
4. 14:02:28 — sidekiq: RewardGiveJob failed, retrying.
5. 14:02:30 — sidekiq WARN: "Queue reward depth above 10,000". This is 78s after the first error (14:02:30 - 14:01:12 = 1m18s).
6. 14:02:36 — first RecognitionDigestJob failure (J-00013); second RewardGiveJob batch J-00007..J-00012 at 14:02:51-14:02:58 (6 more).
7. 14:03:05 — api-gateway: "502 upstream timeout calling reward-service /gives". 113s after first error (14:03:05 - 14:01:12 = 1m53s). First user-facing hop failure.
8. 14:03:30 — web-app: "Give form submission failed: upstream 502 from api-gateway". 138s after first error. End-user impact begins.
9. 14:03:31-14:06:47 — sustained loop: sidekiq RewardGiveJob failures at 14:03:31, 14:04:22, 14:05:26, 14:06:47; api-gateway 502s at 14:03:48, 14:04:13, 14:05:16, 14:06:52; web-app 502s at 14:04:45, 14:05:42, 14:06:49; RecognitionDigestJob failures J-00014 14:03:15, J-00015 14:04:55, J-00016 14:05:50.
10. 14:06:52 — last ERROR (api-gateway 502), 340s = 5m40s after the first error.
11. 14:10:56-14:20:59 — only 6 postgres INFO "checkpoint complete" lines; no errors, unrelated background.
12. 14:22:10 — reward-service INFO: "Redis connection restored; resuming job processing", 1,258s = 20m58s after the first error.
13. 14:24:45 — sidekiq INFO: "Queue reward depth below 500", 155s = 2m35s after restore.

SERVICES AND JOBS
- Origin: reward-service (Redis client timeout). Propagation: sidekiq (job retries + queue backlog) -> api-gateway (502 upstream timeouts to reward-service) -> web-app (give-form submission failures). postgres appears only as INFO checkpoints.
- Job counts from sidekiq_jobs.csv: 16 failures total = 12 RewardGiveJob + 4 RecognitionDigestJob, all Redis::TimeoutError; failure window 14:01:46-14:05:50 = 244s.
- Log-slice error counts: 19 ERROR lines (reward-service 4, sidekiq 6, api-gateway 5, web-app 4) plus 1 WARN (sidekiq queue depth).

DATADOG QUERY TO CONFIRM THE FIRST ERROR
Query: service:reward-service status:error "Redis::TimeoutError" "redis-primary:6379"
Time window: 2026-09-03 14:00:00-14:05:00 UTC, i.e. epoch from=1788444000 to=1788444300 (1788444000 + 300 = 1788444300). This should return the single 14:01:12 connection-timeout line ahead of the "retry exhausted" lines; broadening to service:reward-service status:error over the same window returns 4 lines (the timeout plus 3 retry-exhausted).

WHAT THE LOGS DO NOT SHOW
- Root cause of the Redis timeout: no redis/redis-primary server-side logs exist in the slice; cause (failover, memory, network, overload) is not observable here.
- Cause of recovery: the 14:22:10 "Redis connection restored" INFO names no action, actor, or fix; no remediation or human intervention appears anywhere.
- Final disposition of the 16 failed jobs: no success/completion events after 14:22:10; whether they retried to success is unknown.
- Coverage gap between files: datadog_logs.csv shows 6 sidekiq RewardGiveJob failure lines but sidekiq_jobs.csv has 12 RewardGiveJob rows; the log slice is partial/sampled.
- RecognitionDigestJob failures appear only in sidekiq_jobs.csv; datadog_logs.csv contains no RecognitionDigestJob entries at all.
- 14:06:52-14:22:10: no reward-service or sidekiq logs (only postgres checkpoints), so worker state during the ~15-minute tail is unobserved.
- Actual queue depth value: only the threshold breach "above 10,000" and later "below 500"; no numeric depth.
- Business/user impact: no count of failed gives, affected customers, or lost revenue anywhere in the data.
- postgres is implicated only by absence (checkpoints only), not by any positive check.
feature-flag-state-summaryPASS 1.00
Plain-English question: Summarize what each feature flag controls and who has it on.
Prompt — what the model saw
Summarize the state of the listed feature flags: what each controls per the code excerpt, which companies or segments have it on, targeting rules, and any flag with no code reference or no targets. Company counts must match the export. Do not describe a flag that is not in the export.
cost $0.0140 · 34s · in 1,166 / out 4,042 tok
✓ orphans✓ no_targets✓ counts_match✓ no_invented

✓ Correct Answer (ground truth)

Plain-English answer: Recognition streaks on for 42 beta companies, redeem-flow redesign off with a 12-company target list, plus any flag with no code reference or no targets.
The answer in one line: rule = company counts must equal the export; do not describe flags absent from the export.
code_referencedrecognition_streaks_v2, points_budget_guardrails, slack_dm_nudges, redeem_flow_redesign, analytics_dashboard_v3, ms_teams_app_v2
flags{"recognition_streaks_v2": {"state": "on", "targeting_rules": "segment:beta_companies", "company_count": 42}, "points_budget_guardrails": {"state": "on", "targeting_rules": "all_companies", "company_c
flags_with_no_targetspaused_offboard_cleanup
orphan_flags_no_code_referencelegacy_give_modal, survey_boosters_q3, paused_offboard_cleanup
rulecompany counts must equal the export; do not describe flags absent from the export.
graded checksorphans · no_targets · counts_match · no_invented

Model's actual answer

show the model's full answer
Per-flag state (counts from flags_export.csv, behavior from flag_code.md):

1. recognition_streaks_v2 — on, segment:beta_companies, 42 companies. Controls: gates StreakTracker.record(give) in app/models/recognition.rb, i.e., whether recognition streaks get recorded on a give.

2. points_budget_guardrails — on, all_companies, 220 companies. Controls: gates BudgetService.new(company).enforce!(giver, points) in app/services/budget_service.rb, i.e., points-budget enforcement at give time. This is the only all-companies flag.

3. slack_dm_nudges — on, segment:region_na, 87 companies. Controls: early return in app/jobs/nudge_job.rb unless enabled; when enabled, SlackDm.send_nudge(user) sends the nudge. Nudges therefore flow only for the NA-region segment.

4. redeem_flow_redesign — off, targeted_list, 12 companies. Controls: render path in app/controllers/redeem_controller.rb — RedeemV2Component if on, RedeemV1Component otherwise. State is off, so all companies (including the 12 targeted) render V1 per the excerpt.

5. analytics_dashboard_v3 — on, segment:tier_three, 65 companies. Controls: sets @dashboard = AnalyticsV3.new(company) in app/controllers/analytics_controller.rb. No else branch in the excerpt, so companies outside tier_three get no v3 dashboard assignment from this code.

6. ms_teams_app_v2 — off, targeted_list, 9 companies. Controls: TeamsAppV2.install(company) in app/services/teams_installer.rb. Off, so no v2 installs occur per the excerpt despite 9 companies on the targeted list.

7. legacy_give_modal — off, segment:legacy_plan, 14 companies. No code reference in the excerpt; cannot state what it controls from provided data.

8. survey_boosters_q3 — on, segment:legacy_plan, 7 companies. No code reference in the excerpt. It is on and targeted, but nothing in the provided code consumes it.

9. paused_offboard_cleanup — off, no targeting rule (blank), 0 companies. No code reference and no targets; nothing is enabled anywhere for this flag.

Gaps and anomalies:
- No code reference (3 of 9 flags): legacy_give_modal, survey_boosters_q3, paused_offboard_cleanup.
- No targets: paused_offboard_cleanup (empty targeting_rules, company_count 0).
- Same segment, mismatched counts: survey_boosters_q3 (7) and legacy_give_modal (14) both list segment:legacy_plan. The export gives different counts for the same segment name; the provided data does not explain why.
- Off but targeted: redeem_flow_redesign (12) and ms_teams_app_v2 (9) carry targeted lists while off, so those companies receive no treatment per the excerpt.

Arithmetic (sums of per-flag company_count; companies can overlap across flags, so sums are not unique-company totals):
- On flags (5): 42 + 220 + 87 + 65 + 7 = 421.
- Off flags (4): 12 + 9 + 14 + 0 = 35.
- Total across 9 flags: 421 + 35 = 456.
- Code-referenced flags (6): 42 + 220 + 87 + 12 + 65 + 9 = 435.
- No-code-reference flags (3): 14 + 7 + 0 = 21.
- Cross-check: 435 + 21 = 456, matching the total.
- All segment/targeted counts (42, 87, 65, 14, 7, 12, 9) are ≤ the 220 all_companies count, consistent with subsets of the full customer base.
nda-triagePASS 1.00
Plain-English question: Triage each NDA green, yellow, or red — and quote the clause that decides it.
Prompt — what the model saw
Triage each NDA as GREEN (standard approval), YELLOW (counsel review), or RED (full legal review). Cite the clause that drives each classification by quoting it. Catch embedded non-solicits, non-competes, missing carve-outs, and unusual term or governing law. One line of reasoning per NDA.
cost $0.0068 · 18s · in 1,214 / out 1,651 tok
✓ green✓ yellow✓ red✓ non_solicit_caught✓ clause_quoted✓ correct_labels

✓ Correct Answer (ground truth)

Plain-English answer: NDA-1 green (standard mutual terms), NDA-2 yellow for an embedded 18-month non-solicit, with the driving clause quoted for every classification.
The answer in one line: non_solicit_in = NDA-2
checksclassification_per_nda, clause_quotes_exist_in_document, non_solicit_caught, carveouts_checked
classifications{"NDA-1": {"classification": "GREEN", "driving_reason": "standard mutual NDA: 2-year term, standard carve-outs, Delaware law", "driving_clause_quote": "Carve-outs: information that (a) is or becomes p
non_solicit_inNDA-2
graded checksgreen · yellow · red · non_solicit_caught · clause_quoted · correct_labels

Model's actual answer

show the model's full answer
**NDA-1 ([PARTY A] and Bonusly) — GREEN (standard approval).**
Driving clause: Section 3 carve-outs — "information that (a) is or becomes publicly available through no breach, (b) was known prior to disclosure, (c) is received from a third party without duty of confidence, (d) is independently developed, or (e) must be disclosed by law or court order" — complete set, with no embedded non-solicit/non-compete anywhere in the document.
Reasoning: Mutual, all five standard carve-outs present, Delaware law, and total exposure is 2-year term + 3-year survival = 5 years max obligation tail, within normal bounds — nothing beyond standard approval.

**NDA-2 ([PARTY B] and Bonusly) — YELLOW (counsel review).**
Driving clause: Section 4 — "During the term of this Agreement and for eighteen (18) months thereafter, neither party shall solicit for employment or hire any employee of the other party with whom it came into contact under this Agreement."
Reasoning: An embedded no-hire clause ("or hire"), not just a non-solicit, is out of scope for an NDA and runs 18 months past a 3-year term (36 + 18 = 54 months ≈ 4.5 years of hiring restriction); everything else is clean (mutual, full carve-outs, Delaware law), so this is a counsel-review fix (delete or narrow the clause), not a full legal escalation.

**NDA-3 ([PARTY C] and Bonusly) — RED (full legal review).**
Driving clause: Section 2 — "For a period of three (3) years following the Effective Date, Recipient shall not, directly or indirectly, engage in or provide services to any business that competes with Discloser's business."
Reasoning: A broad, unlimited-geography non-compete bound on Bonusly as Recipient, combined with Section 4's "No confidentiality carve-outs are stated in this Agreement" (even public information stays confidential, which is operationally impossible to honor), Section 3's foreign forum ("laws of the Republic of Ireland... exclusive jurisdiction of its courts"), and Section 1's one-way 5-year term — four material deviations stacked in one document, so full legal review is required before signature.

Missing data note: none of the three files state execution/effective dates, so all term math above is relative to the Effective Date only; no other gaps identified in the provided text.