{"id":"risk-vendor-tier-one-vendor","audience":"compliance","category":"legal-compliance","track":"legal","technique":"constraints","level":"beginner","title":"Propose one vendor tier in three labeled lines","free":false,"task":"Write a prompt that applies the criteria to one vendor and returns three plain-text lines starting Facts:, Criteria met:, Proposed tier:. Under 70 words. The last line must read in the form \"Proposed tier: Tier N - risk owner to confirm\". Count only real personal data.","materialLabel":"Your material (attached to your prompt automatically)","material":"TIERING CRITERIA (Vendor Risk Standard v1)\n- Tier 1: personal data on more than 10,000 people, or can initiate payments or refunds.\n- Tier 2: personal data on 10,000 people or fewer, cannot move money.\n- Tier 3: no personal data.\n\nVENDOR INTAKE - Fernlea Learning (training platform)\n- Will hold names and work emails for 3,000 of our employees.\n- Has a demo area with 40,000 synthetic test learner records (made-up names, not real people).\n- Cannot initiate payments or refunds.\n- Contract value $18,000 a year.\n- The account manager told procurement the platform is \"low risk\".","rules":[{"type":"plainText"},{"type":"maxWords","max":70}],"promptHints":[{"label":"Three labels","patterns":["facts:","criteria met:","proposed tier:"]},{"label":"Excludes synthetic","patterns":["synthetic","test","not real"]},{"label":"Compare to threshold","patterns":["3,000","10,000"]},{"label":"Exact last line","patterns":["risk owner to confirm","last line"]},{"label":"Ignores opinion","patterns":["opinion","account manager"]}],"mistakes":[{"title":"Counting the test records","why":"AI counts 43,000 records and picks Tier 1.","fix":"Say: \"Do not count the 40,000 synthetic records.\"","checks":["r0"]},{"title":"Using the sales view","why":"AI repeats 'low risk' as a reason.","fix":"Say: \"Do not use the account manager's opinion.\"","checks":["r1"]},{"title":"Wrong ending","why":"AI ends with a long recommendation.","fix":"Say: \"Last line exactly: Proposed tier: Tier 2 - risk owner to confirm\"","checks":["r2"]},{"title":"Getting markdown back","why":"AI adds bold, headings and bullets by default, which show up as stray symbols when pasted.","fix":"Say \"plain text, no bold, no headings, no asterisks\".","checks":["plainText"]},{"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":"Give the three labels, say to exclude test data and quote the final line format.","mission":"Propose a tier with the reasoning in three lines.","musts":["Three labeled lines","Synthetic data excluded","Exact tier line"],"rubricLabels":["There are three lines starting Facts:, Criteria met: and Proposed tier: in that order","Facts mentions 3,000 employees and the synthetic records, and Criteria met counts the 3,000 but not the 40,000 synthetic records","The last line reads exactly: Proposed tier: Tier 2 - risk owner to confirm, and the account manager's \"low risk\" remark is not used as a reason"]}