Tasks / Operate

Make the launch call

Can the model make a clear go / no-go call from mixed launch evidence, checked against the criteria agreed up front?

Measures the modelTask v1.2 · 2 casesDifficulty

What AI gets right here, and what you’ll still have to catch

From 12 graded outputs by 6 models. 75% were usable with at most a quick edit.

Reliably right

Finds the Scotland breach100% pass
Clearly identifies Scotland's refund breach (66% above control, 143% after supplier change) and acceptance dip below 80%, correctly holds Scotland.
Opus 5.5 · Claude · Go/no-go for AI substitutions, from the rollout data
Respects explicit constraints98% pass
The output is a launch recommendation, addresses the meeting, and stays within the 400-word limit (345 words).
GPT-6 Astra · ChatGPT · Go/no-go for AI-drafted support replies
Produces the required deliverable98% pass
The recommendation is in the requested form, for the meeting, within the length, and includes necessary steps a PM could act on.
GPT-6 Astra · ChatGPT · Go/no-go for AI-drafted support replies

Where it slips

Limits the downside of being wrong35% pass
No specific post-launch monitoring signal or threshold is named; only generic 'monitoring' and a kill switch are mentioned.
GPT-6 Astra · ChatGPT · Go/no-go for AI-drafted support replies
Catches the duplicate rows54% pass
Finds and removes the duplicate rows but never states whether they change the results.
GPT-6.1 Sol · API · Go/no-go for AI substitutions, from the rollout data
Gets the base of every number right67% pass
Some derived figures are off, including the pooled basket gap and the control-order share for London, and the basket base is stated as means of daily averages rather than order-weighted averages.
Sonnet 5.5 · API · Go/no-go for AI substitutions, from the rollout data

Case viewer

Read the brief, then put up to three outputs side by side, each with the LLM judge’s verdict on every check. Highlights mark what a PM had to fix.

The brief

You're the Staff PM for fulfilment at Basketful. On 30 October the launch meeting decides whether AI auto-substitutions (the model picks a replacement when an item is out of stock, instead of the picker) roll out to every region on Monday 2 November, ahead of the Christmas delivery-slot booking window that opens on 9 November. The rollout workbook is attached as two files: daily_metrics.csv (28 days of the regional test, by region and arm) and incidents.csv. Work from the data, not the summaries people have given you. Write your recommendation for the meeting: go, no-go or go with conditions, region by region, and why. Include a table that checks each launch criterion for each region with the figures. Keep the prose under 900 words.

The test1–28 October, five regions. In each region orders were split between control (the picker chooses substitutes, as today) and treatment (AI auto-substitutions). The split was not 50/50 everywhere: treatment got 30% of orders in London, 50% in the South East and North West, and 70% in the Midlands and Scotland.

Launch criteria (agreed in the PRD)1. Substitution acceptance (subs_accepted ÷ subs_offered) of at least 80% in treatment, in every region. 2. Refund requests per 1,000 orders in treatment no more than 10% above control, in every region. 3. Zero dietary or allergen mismatches (a substitute that breaks a gluten-free, vegan, vegetarian, nut-free, halal or kosher attribute). 4. Average basket value not lower in treatment than control.

What people have saidHead of Commercial: “Acceptance is up 15 points and refunds are within the guardrail overall. Every day we wait costs us Christmas.” Head of Analytics: “Treatment baskets are £4 smaller across the test. That worries me.” Ops director (Scotland): “The supplier change caused some noise, but it's settling down.”

EngineeringThe dietary-attribute hard filter (never substitute across those attributes) is merged and scheduled to deploy on 4 November after regression tests. Scotland's catalogue re-map for the new dairy supplier is in progress with no date. Rollout can be switched on region by region.

daily_metrics.csv283 rows · Download

date,region,arm,orders,items_ordered,items_out_of_stock,subs_offered,subs_accepted,refund_requests,refund_value_gbp,complaints,avg_basket_gbp
2026-10-01,London,control,1757,52273,2351,1953,1387,32,160.71,6,72.89
2026-10-01,London,treatment,692,20455,871,820,708,12,62.05,2,72.42
2026-10-01,South East,control,853,22846,996,833,593,16,74.49,3,65.62
2026-10-01,South East,treatment,928,25223,1166,1106,952,15,75.93,3,66.59
2026-10-01,Midlands,control,467,10468,465,383,268,8,34.57,1,54.92
2026-10-01,Midlands,treatment,1113,25136,1166,1087,947,21,106.35,3,55.33
…

incidents.csv13 rows · Download

date,region,arm,type,severity,summary,status
2026-10-02,London,treatment,wrong_size,low,"4-pint milk substituted with 1-pint, customer accepted at door",closed
2026-10-04,Midlands,treatment,dietary_mismatch,high,"Gluten-free sliced loaf substituted with standard white loaf; customer coeliac, noticed at home",closed: refund and apology
2026-10-06,South East,control,wrong_item,low,"Picker substituted oat milk with soya milk; customer rejected at door",closed
2026-10-07,North West,treatment,data_export,low,"Daily export job re-ran after a timeout on 8, 9 and 10 Oct; analytics team flagged possible duplicate rows",open
2026-10-09,London,treatment,price,medium,"Substitute priced higher than original; customer charged the difference against policy",closed: policy fix deployed 11 Oct
2026-10-12,Scotland,control,late_delivery,low,"Van breakdown, 14 orders late",closed
…

What a strong answer does

No-go for a national launch on 2 November, because criterion 3 fails everywhere until the dietary hard filter ships: incidents.csv has three treatment dietary mismatches (4 Oct Midlands, 19 Oct South East, 23 Oct London, the last involving nuts and a child with an allergy). Go with conditions, region by region, once the filter is deployed and verified (planned 4 November, still before the 9 November booking window): London, South East, Midlands and North West pass the other criteria (acceptance about 86% against 71% control; refunds 2–5% below control). Hold Scotland: over the month its treatment refunds are about 66% above control, and from the 15 October supplier change about 143% above (43 against 18 per 1,000 orders), with acceptance falling to about 79%, below the 80% bar; it waits for the catalogue re-map and a clean re-test. The pooled refund rate (about 8% above control) hides this, and the criterion is per region. The pooled basket gap (about £65 against £61) is an artefact of the uneven split, since treatment baskets are equal or higher in every region (Simpson's paradox), so criterion 4 passes. Notices that North West treatment rows for 8–10 October are duplicated (an extra 2,194 orders), removes them, and says it doesn't change the result. States what would change the call: the filter slipping past 9 November, or Scotland's refunds not recovering.

Critical failures (cap the score)

Recommends any rollout before the dietary hard filter is deployed
Recommends rolling out to Scotland while its refund guardrail is breached

Case

v1.6 · synthetic · e-commerce, AI product, data analysis, staff-level

Opus 5.5ClaudeUsable with a quick edit Sonnet 5.5APIUsable with a quick edit GPT-6.1 SolAPIUsable with a quick edit GPT-6 AstraChatGPTUsable with a quick edit GPT-6 LunaAPINeeds more work Gemini 3.5 Flash-LiteGeminiNeeds more work

GPT-6.1 SolAPI

Usable with a quick edit10 of 13 checks passed

The recommendation, table, and calculations are solid; only minor wording additions to state duplicate impact and flag irreversible dietary harm are needed.

Output

Recommendation for 30 October: go with conditions in four regions; no-go in Scotland

Do not expand any region on Monday 2 November. London, South East, Midlands and North West are candidates for a gated rollout after the dietary hard filter is deployed and validated—not before its scheduled 4 November deployment. Scotland must additionally resolve and re-test its catalogue mapping.

Christmas urgency does not justify launching ahead of a known safety fix. The 9 November booking window should guide sequencing, not override the PRD.

Method and launch-criterion checks

I removed three exact duplicate North West treatment rows, dated 8–10 October, consistent with the export incident. This leaves 280 unique date–region–arm records.

Acceptance and refunds below use ratios of summed counts, not averages of daily rates. Basket values are order-weighted. T = treatment; C = control. Dietary counts are recorded treatment incidents; zero recorded incidents is not proof that none occurred.

Region	Treatment acceptance ≥80%	Refunds/1,000 orders: T vs C; relative change ≤+10%	Zero dietary/allergen mismatches	Average basket: T vs C; difference ≥£0	Recommendation
London	Pass: 86.7% (23,607/27,234)	Pass: 17.44 vs 18.04; −3.3%	Fail: 1, nut-free → hazelnuts, 23 Oct	Pass: £72.70 vs £72.49; +£0.21	Go with conditions; not 2 Nov
South East	Pass: 86.4% (26,586/30,770)	Pass: 17.54 vs 17.89; −1.9%	Fail: 1, vegan → dairy, 19 Oct	Pass: £66.64 vs £66.20; +£0.44	Go with conditions; not 2 Nov
Midlands	Pass: 86.3% (26,732/30,958)	Pass: 17.48 vs 17.89; −2.3%	Fail: 1, gluten-free → standard bread, 4 Oct	Pass: £55.66 vs £55.35; +£0.31	Go with conditions; not 2 Nov
North West	Pass: 86.3% (18,153/21,039)	Pass: 17.58 vs 18.47; −4.8%	Pass on incident log: 0 recorded	Pass: £57.51 vs £57.29; +£0.22	Go with conditions; not 2 Nov
Scotland	Pass over 28 days: 82.8% (14,333/17,319)	Fail: 30.16 vs 18.20; +65.7%	Pass on incident log: 0 recorded	Pass: £52.98 vs £52.62; +£0.36	No-go

What the data actually say

Commercial’s overall refund claim hides a regional failure. After deduplication, pooled refunds are 19.50 per 1,000 treatment orders versus 18.08 control: +7.9%, within the overall guardrail. But the PRD requires compliance in every region, and Scotland fails badly. Strong acceptance cannot compensate for dietary breaches.

Analytics’ £4 basket concern is real in the pooled arithmetic, but not within regions. Pooled baskets are £61.09 treatment versus £65.15 control (−£4.06). Treatment disproportionately includes lower-basket Midlands and Scotland orders, while control disproportionately includes higher-basket London orders. Every within-region comparison is positive. This is a composition effect, not evidence here that AI reduces basket value. These are point estimates; the daily averages alone do not establish statistical certainty.

Scotland is not settling down. Following the supplier switch on 15 October:

Acceptance falls from 86.3% on 1–14 October to 79.1% on 15–28 October, below the launch threshold.
Treatment refunds rise from 17.25 to 43.25 per 1,000 orders. Post-switch control is 17.79, making treatment 143.1% worse.
The final week still has 79.2% acceptance and 43.64 refunds per 1,000.

The open pack-size mapping incident is consistent with this sustained deterioration. The full-period acceptance pass masks the current operating state.

Conditions and release plan

For London, South East, Midlands and North West, require:

Deployed, validated safety protection: successful regression tests across gluten-free, vegan, vegetarian, nut-free, halal and kosher attributes, including missing/incorrect metadata and the three observed failures. “Merged” is not “protected.”
Fresh post-fix evidence: a controlled validation window with zero dietary mismatches and all regional numerical criteria met, with explicit launch sign-off. Refunds and apologies do not erase historical criterion failures.
Staged regional expansion: monitor acceptance, refunds and dietary incidents daily; stop immediately for any dietary mismatch and roll back for guardrail breaches. Reconcile the North West export before reporting.

For Scotland, also require a completed catalogue re-map, pack-size validation and fresh post-remediation evidence. With no completion date and continuing deterioration, make no Christmas rollout commitment.

Meeting decision: reject the all-region 2 November launch; approve a gated path for four regions, while keeping Scotland blocked.

Check by check

Got wrong · 2

Catches the duplicate rowsFinds and removes the duplicate rows but never states whether they change the results.
Limits the downside of being wrongGives monitoring and rollback thresholds but does not flag irreversible consequences such as customer-facing dietary substitutions or communications.

Mixed · 1

Gets the base of every number rightAll percentages, per-1,000 rates, and basket differences use correct summed or order-weighted bases.The two graders disagreed on this one.

Got right · 10

Uses the supplied evidence correctlyAll current-state numbers and facts trace to the brief, incidents.csv, or arithmetic from daily_metrics.csv.
Addresses the actual decisionCommits to a clear region-by-region go/no-go with dates and conditions for the launch meeting.
Respects explicit constraintsIncludes the required region-by-criterion table and keeps prose within the length while respecting the no-early-launch safety constraint.
Identifies material uncertaintyNames filter deployment/validation and Scotland catalogue remap as unresolved and ties go decisions to fresh evidence.
Avoids unsupported claimsCausal claims are hedged ('consistent with') and interpretations are labelled as composition effects or point estimates.
Produces the required deliverableProvides a usable recommendation, table, and release conditions that the reader could act on with light edits.
Checks every criterion, region by regionRegion-by-criterion table covers acceptance, refunds vs control, dietary incidents, and basket values with computed figures.
Finds the Scotland breachFinds Scotland's refund breach, ties it to the supplier change with post-15 Oct numbers, and holds Scotland.
Explains the basket gapExplains the pooled basket gap as a traffic-mix effect and shows all within-region treatment baskets positive.
Makes a clear callMakes a clear per-region call with dates and conditions, and Scotland is blocked.

Grades and run details

Decision model 85 · LLM judge 12 of 14 checks

Decision model checks

passUses the supplied evidence correctly26%
passAddresses the actual decision81%
passRespects explicit constraints41%
passIdentifies material uncertainty46%
passAvoids unsupported claims42%
passProduces the required deliverable79%
passChecks every criterion, region by region71%
passFinds the Scotland breach100%
passExplains the basket gap99%
partialCatches the duplicate rows49%
passMakes a clear call34%
partialLimits the downside of being wrong43%
failGets the base of every number right22%

Run

Run: #1
API response time: 1.9 min
Submitted: 29 Sept 2026

Sonnet 5.5API

Usable with a quick edit11 of 13 checks passed

A PM could use it with light edits, mainly correcting minor arithmetic and removing the unsupported incident-reporting inference. The core decision, table, and conditions are actionable.

Output

Recommendation: AI auto-substitutions, launch meeting 30 October

Decision: no-go for Monday 2 November. Go with conditions from 4 November for London, South East, Midlands and North West. Scotland stays on picker substitutions until it re-passes.

Criteria by region

Figures are recomputed from the daily rows. Treatment (T) and control (C) are pooled over the 28 days.

Region	1. Acceptance (T) ≥80%	2. Refunds per 1,000 orders, T vs C (≤ +10%)	3. Dietary mismatches (0)	4. Basket T vs C (not lower)	Verdict
London	86.7% ✅	17.4 vs 18.0 (−3%) ✅	1 (nut allergen, 23 Oct, open) ❌	£72.72 vs £72.48 (+£0.24) ✅	Conditional
South East	86.4% ✅	17.5 vs 17.9 (−2%) ✅	1 (vegan cheese, 19 Oct) ❌	£66.66 vs £66.20 (+£0.46) ✅	Conditional
Midlands	86.3% ✅	17.5 vs 17.9 (−2%) ✅	1 (gluten-free, 4 Oct) ❌	£55.66 vs £55.35 (+£0.31) ✅	Conditional
North West	86.3% ✅	17.6 vs 18.5 (−5%) ✅	0 logged ✅	£57.52 vs £57.28 (+£0.24) ✅	Conditional
Scotland	82.8% overall, 79.1% since 15 Oct ❌	30.2 vs 18.2 (+66%); since 15 Oct 43.3 vs 17.8 (+143%) ❌	0 logged ✅	£52.98 vs £52.62 (+£0.36) ✅	No-go

Baskets are means of daily averages. The differences are small and I have not tested them for significance, but the point estimate is positive everywhere.

Checking what people have said

Head of Commercial. The pooled numbers are right: acceptance is 85.9% against 71.2% (+14.7 points), and refunds are 19.5 against 18.1 per 1,000 (+8%). But the pooled refund figure hides Scotland, which is failing badly. Four regions are 2–5% better than control on refunds.
Head of Analytics. The £4 gap is real in the pooled data (£61.10 vs £65.15) but it is a mix effect, not a treatment effect. London, with the highest baskets (about £72), was 42% of control orders but only 18% of treatment orders. The Midlands and Scotland (about £53–56) were 18% of control but 42% of treatment. Within every region, treatment baskets are £0.24–£0.46 higher. Criterion 4 passes. The uneven split did not bias the regional comparisons.
Ops director (Scotland). The data does not support "settling down". Before the supplier switch on 15 October, Scotland treatment was healthy (86.3% acceptance, 17.3 refunds per 1,000). From 15 October to 28 October, acceptance has been flat at 78–81%, and daily refunds have been 25–33 against a pre-switch level of 10–14. The cause is the open incident: the new supplier's pack sizes are mapped to the wrong product family, so the model makes wrong-size dairy substitutions. That is a catalogue defect with no fix date, not noise. Scotland control is unaffected.

Why the date moves

Criterion 3 fails in three regions. The three mismatches include a nut-allergy case and a coeliac case. The hard filter that prevents them deploys on 4 November. Until it is live, every region has the same exposure. North West and Scotland show zero only because none were reported, and the incidents come from customer complaints, so the true rate is unknown. I would not switch on a known allergen risk two days before the fix.

Waiting is cheap. The booking window opens on 9 November, not 2 November. Deploying the filter on 4 November and switching regions on 4–5 November leaves several days before the window. Only Scotland has no route to launch before then.

Conditions for the four regions

The filter deploys on 4 November and passes regression tests. Before any region goes live, it is tested in production against the three incident SKUs (gluten-free loaf, vegan cheese, nut-free bar) and the other attributes: vegetarian, halal and kosher. The London incident is still open with the filter "in review", so confirm the filter would have blocked it.
Regions are switched on region by region, North West and South East first, then London and the Midlands. This gives a day of live data before the larger volumes go on.
Any dietary or allergen mismatch after launch switches that region back to picker substitution immediately, pending review. That includes any mismatch in the first week.
Daily monitoring of acceptance (80% floor) and refunds (10% guardrail) per region against a small held-back control. Delegate the switch-on decision so it does not need another meeting.

Scotland

Keep Scotland on picker substitutions. Retest only after the catalogue re-map ships, on a small treatment share. It should show at least three consecutive days at 80% or better acceptance, and refunds within 10% of control, before it goes to full volume. Engineering has no date, so Scotland should not be in the Christmas plan and should be treated as upside.

Data notes

North West treatment has three duplicated rows (8, 9 and 10 October), matching the open export-job incident. I removed them. This restores the intended 50% split, and the North West rates barely move either way.
Incident logs are a floor on dietary failures, not a measure of them.
This is 28 days of data in a low-volume region (Scotland control is about 265 orders a day). Even so, the Scotland gap is far outside noise.

Check by check

Got wrong · 1

Gets the base of every number rightSome derived figures are off, including the pooled basket gap and the control-order share for London, and the basket base is stated as means of daily averages rather than order-weighted averages.

Mixed · 1

Avoids unsupported claimsIt asserts that zero logged dietary incidents in North West and Scotland are only because none were reported and that incidents come from customer complaints, which is not established by the supplied evidence.The two graders disagreed on this one.

Got right · 11

Uses the supplied evidence correctlyThe output's factual claims are drawn from the brief and supplied CSVs, with only minor rounding and one unsupported inference about incident reporting.
Addresses the actual decisionIt commits early to no-go on 2 November, conditional go from 4 November for four regions, hold Scotland, and states what would change the call.
Respects explicit constraintsIt provides a region-by-region recommendation with a launch-criteria table and keeps prose under 900 words.
Identifies material uncertaintyIt names material unknowns such as filter deployment, Scotland re-map timing, duplicate handling, and incident under-reporting, and gives monitoring/rollback conditions.
Produces the required deliverableThe requested recommendation and table are present, usable for the launch meeting, and within the length limit.
Checks every criterion, region by regionIt checks all four launch criteria for each region with figures computed from the files, including the duplicate correction.
Finds the Scotland breachIt identifies Scotland's refund breach, ties it to the 15 October supplier change, notes the acceptance dip, and holds Scotland.
Explains the basket gapIt compares baskets within each region and explains the pooled gap as an artefact of the uneven regional split.
Catches the duplicate rowsIt finds the duplicated North West treatment rows, removes them, and says the result barely changes.
Makes a clear callIt gives one clear call per region with dates, conditions, and triggers for change.
Limits the downside of being wrongIt limits downside with staged switch-on, daily monitoring, immediate rollback on any dietary mismatch, and Scotland held out.

Grades and run details

Decision model 92 · LLM judge 12 of 14 checks

Decision model checks

passUses the supplied evidence correctly14%
passAddresses the actual decision99%
passRespects explicit constraints29%
passIdentifies material uncertainty90%
passAvoids unsupported claims23%
passProduces the required deliverable80%
passChecks every criterion, region by region55%
passFinds the Scotland breach100%
passExplains the basket gap99%
passCatches the duplicate rows100%
passMakes a clear call84%
passLimits the downside of being wrong72%
failGets the base of every number right36%

Run

Run: #1
API response time: 2.3 min
Submitted: 29 Sept 2026

Results

Every setup we’ve tested on this task, across all cases and repeats, graded on the current checklist. Calibrated: the graders match our PM on 86% of checks.

#	Model · Harness	Decision model	LLM judge	Runs	Critical failures
1	Sonnet 5.5withAPI	93.7	88.3	2	None
2	GPT-6 AstrawithChatGPT	89.8	88.3	2	None
3	GPT-6.1 SolwithAPI	87.3	88.3	2	None
4	Opus 5.5withClaude	80.4	86.4	2	None
5	GPT-6 LunawithAPI	86.0	77.6	2	None
6	Gemini 3.5 Flash-LitewithGemini	53.1	42.2	2	None

The PM job

Deciding whether a feature ships on the planned date, and on what conditions.

Why it matters

Launch meetings reward optimism. A good call checks each agreed criterion, weighs a real risk against a real gain, and says exactly what would change the answer. A weak one rubber-stamps the launch or blocks it on noise.

What good looks like

Checks each agreed criterion against the evidence
Makes one clear call, with conditions if needed
Separates risks that block launch from risks that can be managed
Says what would change the call
Says how to limit the damage if the call is wrong

Deliberately not measured

Rollout engineering detail
Project-plan formatting

Capability tested

Deciding against agreed launch criteria

The failure we’re looking for

Rubber-stamps a launch that misses an agreed bar, or blocks it on noise

Grading

Decision model and LLM judge, calibrated against a blind PM review

Variants

A launch memo from supplied evidence · Staff level: a regional call from an attached data workbook

What AI gets right here, and what you’ll still have to catch

Reliably right

Where it slips

Case viewer

Output

Recommendation for 30 October: go with conditions in four regions; no-go in Scotland

Method and launch-criterion checks

What the data actually say

Conditions and release plan

Check by check

Got wrong · 2

Mixed · 1

Got right · 10

Grades and run details

Decision model checks

Run

Output

Recommendation: AI auto-substitutions, launch meeting 30 October

Criteria by region

Checking what people have said

Why the date moves

Conditions for the four regions

Scotland

Data notes

Check by check

Got wrong · 1

Mixed · 1

Got right · 11

Grades and run details

Decision model checks

Run

Results

About the task

The PM job

Why it matters

What good looks like

Deliberately not measured