{"id":"ind-np-grant-outcomes-met","audience":"nonprofit","category":"fact-checking","technique":"verification","title":"Report grant outcomes honestly as Met or Not met","free":false,"task":"Your grant report asks whether each promised outcome was met. Write exactly 3 numbered lines in the form \"Outcome: target X, actual Y - Met or Not met\", using only those two labels, plain text, under 60 words. Use the final evaluation numbers and report a miss as a miss, with no excuses or spin on the line. Write a prompt that gets you exactly that.","materialLabel":"The grant targets and evaluation (attached to your prompt automatically)","material":"GRANT OUTCOMES - \"Bright Futures Mentoring\" - funded by the Calloway Foundation (fictional)\n\nPromised outcomes (from the proposal):\n1. Match 80 youth with adult mentors.\n2. 75% of matches last at least 9 months.\n3. 70% of youth report higher confidence on the end-of-year survey.\n\nProgram staff draft (Oct 1): 84 matched; 78% lasted 9+ months; 72% higher confidence.\nFINAL EVALUATION (outside evaluator, Nov 3): 84 matched; 71% lasted 9+ months (some matches ended in month 8 when families moved); 72% reported higher confidence.\n\nNotes:\n- The foundation's form only allows \"Met\" or \"Not met.\"\n- Executive director wants to \"frame the 71% positively.\" The form has a separate narrative section for context.","rules":[{"type":"numberedItems","count":3},{"type":"maxWords","max":60},{"type":"plainText"}],"mistakes":[{"title":"No fixed shape","why":"The AI writes a narrative report with context for each outcome.","fix":"Be exact: \"Exactly 3 numbered lines: Outcome: target X, actual Y - Met or Not met.\"","checks":["numberedItems","plainText"]},{"title":"Trusting the AI to count words","why":"AI is bad at counting. Ask for \"under 60 words\" and it often lands well over.","fix":"Cap each part instead: \"each line under 15 words.\"","checks":["maxWords"]},{"title":"Using draft numbers","why":"The staff draft's 78% looks better and appears first.","fix":"Add: \"Use the final evaluation (Nov 3).\"","checks":["r0"]},{"title":"Softening the miss","why":"The AI writes \"Nearly met\" or adds why it fell short.","fix":"Add: \"Use only Met or Not met. No reasons on the line.\"","checks":["r1","r2"]}],"promptHints":[{"label":"Uses final numbers","patterns":["final","evaluat","nov(ember)? 3","71"]},{"label":"Uses exact labels","patterns":["met or not met","only (those|these|two) labels","exact(ly)? (labels|words)"]},{"label":"Rules out spin","patterns":["no (excuses|spin|framing)","honest","don'?t explain"]},{"label":"Shows the line shape","patterns":["target x","actual y","outcome:"]},{"label":"Sets a word limit","patterns":["\\d+\\s*words","word (limit|count)"]}],"hint":"Funders trust honest reporting. Tell the AI which numbers are final, the only two labels allowed, and to keep reasons out of the line.","rubricLabels":["Uses the final evaluation numbers: 84 matched (target 80), 71% lasted 9+ months (target 75%), 72% higher confidence (target 70%), never the draft's 78%","Labels are exactly Met, Not met, Met for outcomes 1, 2 and 3 respectively, with no other labels (like Partially met or Nearly met)","Each line follows 'Outcome: target X, actual Y - label' with no excuses, reasons or positive framing (nothing about families moving or 'close to target')"]}