{"id":"res-pm-prd-claims-interview-evidence","audience":"product-managers","category":"fact-checking","technique":"verification","title":"Check your PRD's user claims against the interview notes","free":false,"task":"Your PRD for a team calendar feature makes 4 claims about users, and you want to check them against your interview notes before review. You want exactly 4 numbered lines, one per claim, labeled Supported or Not supported, with the interview IDs that back each Supported claim in parentheses. A claim needs at least 2 interviews to count as Supported. Plain text, nothing else.","materialLabel":"Your interview notes and PRD claims (attached to your prompt automatically)","material":"USER INTERVIEWS - Team calendar research (Oct 2026)\nU1 (ops manager): \"I keep a spreadsheet of who's out because our tool doesn't show it.\"\nU2 (team lead): \"Seeing PTO next to deadlines would save me so many surprises.\"\nU3 (designer): \"Honestly I just want dark mode.\"\nU4 (team lead): \"I have to ask in Slack who's out every Monday.\"\nU5 (engineer): \"Calendar sync with Google would be nice but not a big deal.\"\nU6 (ops manager): \"If it showed time zones for remote teammates, that would help scheduling.\"\nU7 (team lead): \"Deadlines and out-of-office in one view, yes please.\"\n\n-----\nPRD CLAIMS (draft v1)\n1. Managers struggle to see who is out of office.\n2. Users want PTO shown alongside project deadlines.\n3. Google Calendar sync is a top request.\n4. Users need time-zone display for remote teams.","rules":[{"type":"numberedItems","count":4},{"type":"plainText"}],"mistakes":[{"title":"No bar for 'supported'","why":"The AI counts one mention as evidence, so single requests look like user needs.","fix":"Say: \"Supported means at least 2 interviews back it.\"","checks":["r1"]},{"title":"Not asking for IDs","why":"Without interview IDs, reviewers can't check your evidence.","fix":"Add: \"List the interview IDs for each Supported claim.\"","checks":["r0"]},{"title":"Getting product advice","why":"The AI starts prioritizing features instead of checking claims.","fix":"Say: \"Output only the 4 lines.\"","checks":["r2"]},{"title":"Not saying how many","why":"The AI decides the length itself and usually returns more lines, one per claim than you can use.","fix":"Be exact: \"exactly 4 numbered lines, one per claim.\"","checks":["numberedItems"]},{"title":"Getting markdown back","why":"AI defaults to bold labels and headings, which show up as stray symbols in the PRD comments.","fix":"Ask for \"plain text, no bold, no headings.\"","checks":["plainText"]}],"promptHints":[{"label":"Sets the evidence bar","patterns":["at least (2|two)","more than one","two or more"]},{"label":"Asks for interview IDs","patterns":["\\bIDs?\\b","U\\d","which interviews"]},{"label":"Gives exact labels","patterns":["supported","not supported","label"]},{"label":"Says exactly how many","patterns":["\\b(4|four)\\b.{0,25}(line|claim)","exactly \\d+"]},{"label":"Asks for plain text","patterns":["plain","no bold","no markdown","no formatting","no headings"]}],"hint":"One user saying something isn't a pattern. Give the AI a clear evidence bar and ask it to show which interviews back each claim.","rubricLabels":["Labels claim 1 Supported with U1 and U4 (and optionally U2 or U7), and claim 2 Supported with U2 and U7","Labels claim 3 Not supported (only U5, who said it's not a big deal) and claim 4 Not supported (only U6 mentions it)","Outputs only the 4 labeled lines with interview IDs on Supported lines: no intro, no rewritten PRD, no roadmap advice"]}