{"id":"risk-claims-new-formula-conflict","audience":"compliance","category":"legal-compliance","track":"legal","technique":"verification","level":"advanced","title":"Reconcile claims with test results from the old and new product","free":false,"task":"The product page describes the current v2 bottle, but some evidence in the notes covers v1. Write a prompt for a markdown table with columns Claim | Evidence for v2 | Status | Question, one row per claim C1 to C4, under 190 words. Status is only Supported, Not supported or Needs review. Use v2 evidence only; if there is none, say so. The last line must read: Open for reviewer: followed by the ids that are not Supported, in order.","materialLabel":"Your material (attached to your prompt automatically)","material":"PRODUCT PAGE - TrailPro insulated bottle (v2, on sale since 1 October)\nC1: Keeps drinks cold for 24 hours\nC2: BPA-free\nC3: Dishwasher safe\nC4: Made from 90% recycled steel\n\nEVIDENCE FOLDER\n- Cold test: v1 bottle held under 6C for 24 hours (lab report, 2023). v2 has a new lid; Sept test of v2: 19 hours.\n- BPA: supplier certificate covers the v1 lid. For the v2 lid the supplier wrote \"certificate pending, expected December\".\n- Dishwasher: v2 bottle and lid passed 100 cycles in the internal test, Sept.\n- Recycled steel: supplier invoice for the v2 batch states 70% recycled content. Sourcing lead email: \"90% from the next batch\".\n- Page copy was signed off by Marketing on 28 September.","rules":[{"type":"maxWords","max":190}],"promptHints":[{"label":"Restricts to v2","patterns":["v2","ignore v1","only v2"]},{"label":"Fixes the statuses","patterns":["not supported","needs review"]},{"label":"Sets the table columns","patterns":["columns?","table"]},{"label":"Closing line","patterns":["open for reviewer","last line"]},{"label":"Caps cell length","patterns":["under \\d+","\\d+\\s*words"]}],"mistakes":[{"title":"Using the v1 result","why":"AI cites the 24 hour v1 test as support for the v2 page.","fix":"Say: \"Use v2 evidence only; ignore v1 results.\"","checks":["r0"]},{"title":"Wrong status for pending","why":"AI marks a pending certificate as Supported or Not supported.","fix":"Say: \"Needs review when evidence is pending or partial.\"","checks":["r1"]},{"title":"Missing the closing line","why":"AI ends with its own summary or recommendation.","fix":"Say: \"Last line exactly: Open for reviewer: then the ids not Supported.\"","checks":["r2"]},{"title":"Trusting the AI to count words","why":"AI is bad at counting and usually runs long, so a vague \"keep it short\" lands over the limit.","fix":"Cap each part instead of the total, for example \"each line under 15 words\".","checks":["maxWords"]}],"hint":"Tell it to ignore v1 results and give the exact closing line format.","mission":"Check four claims using only the v2 evidence.","musts":["v2 evidence only","Fixed three statuses","Open-ids closing line"],"rubricLabels":["The table has columns Claim, Evidence for v2, Status and Question with four rows C1 to C4, and every Status is exactly Supported, Not supported or Needs review","C1 and C4 are Not supported using the v2 figures (19 hours, 70%), C2 is Needs review because the v2 lid certificate is pending, and C3 is Supported","The last line reads exactly: Open for reviewer: C1, C2, C4, and the table does not use the v1 24 hour result as evidence for C1"]}