{"id":"eval-mkt-error-analysis-promo-email-drafts","audience":"marketing","category":"evaluation","track":"evals","technique":"format","title":"Sort 13 flagged AI promo emails into 3 failure types","free":false,"task":"Your team flagged 13 AI-drafted promo emails and noted what was wrong with each. Group them into exactly 3 numbered failure types, each line 'Type: #ids = count', then one last line 'Total: N'. Leave out any flag that was withdrawn. Plain text, under 70 words.","materialLabel":"The review notes on flagged emails (attached to your prompt automatically)","material":"Juniper & Pine Candles - AI promo email review, Nov 1-5\nOne note per flagged draft:\n\n#1 says 25% off, the promo is 20%\n#2 link goes to the old summer sale page\n#3 \"the best candles in the universe\" - hype, off brand\n#4 says sale ends Nov 30, it ends Nov 15\n#5 button has no link at all\n#6 \"insanely amazing scents!!!\" - not our voice\n#7 promo code written as FALL25, real code is FALL20\n#8 flag withdrawn by Mel: the free-shipping line was correct after all\n#9 unsubscribe link points to a 404\n#10 \"you'd be crazy to miss this\" - pushy, off brand\n#11 says free shipping over $25, it's over $50\n#12 product link goes to a discontinued candle\n#13 \"literally life-changing\" - hype\n\nContext: we send 3 promo emails a week. Mel wants to know which problem to fix in the prompt first.","rules":[{"type":"numberedItems","count":3},{"type":"plainText"},{"type":"maxWords","max":70}],"mistakes":[{"title":"Too many types","why":"The AI splits discount, date and code errors into separate groups, so no fix stands out.","fix":"Say: \"Exactly 3 failure types.\"","checks":["numberedItems","r0"]},{"title":"No ids or counts","why":"The AI describes problems in prose without showing which drafts belong where.","fix":"Show the shape: \"Type: #1, #4 = 2\".","checks":["r2"]},{"title":"Counting a withdrawn flag","why":"The AI counts every note, including the one Mel withdrew.","fix":"Add: \"Leave out withdrawn flags, and end with Total: N.\"","checks":["r1"]},{"title":"Getting markdown back","why":"AI defaults to bold headings and long bullet lists.","fix":"Ask for \"plain text, no bold, no headings.\"","checks":["plainText","maxWords"]}],"promptHints":[{"label":"Says exactly how many","patterns":["\\b(3|three)\\b.{0,25}(type|group|categor)","exactly \\d+"]},{"label":"Asks for ids and counts","patterns":["#ids","\\bids\\b","count","how many"]},{"label":"Handles the withdrawn flag","patterns":["withdr","#8","leave out","exclude"]},{"label":"Asks for a total","patterns":["total"]},{"label":"Asks for plain text","patterns":["plain","no bold","no markdown","no formatting","no headings"]}],"hint":"Grouping flagged drafts shows which prompt fix pays off most. Fix the number of types, ask for the ids behind each count, and drop withdrawn flags.","rubricLabels":["Uses these 3 groups (names may differ): wrong offer details (discount, date, code or shipping threshold) #1, #4, #7, #11 = 4; broken or wrong link #2, #5, #9, #12 = 4; hype or off-brand wording #3, #6, #10, #13 = 4","Leaves out #8 and ends with a line saying Total: 12","Each type line lists its # ids followed by the count, and each count equals the number of ids on that line"]}