{"id":"search-mkt-two-answers-holiday-survey","audience":"marketing","category":"ai-search","track":"research","technique":"grounding","title":"Check two AI answers about a shopping survey before you quote it","free":false,"task":"You want one stat from the Fenwick holiday shopping survey for a campaign email, and two AI tools described it differently. Write exactly 4 numbered lines labeled A error:, B error:, Safe stat: and Verdict:, judged only against the survey summary. Safe stat: must be one sentence you could paste into the email, with the exact number and the survey name. Plain text, under 140 words.","materialLabel":"The two AI answers and the survey summary (attached to your prompt automatically)","material":"ANSWER A (Tool 1)\n\"The Fenwick 2026 Holiday Survey found 72% of Americans plan to shop mostly online this year, up from 50% last year.\"\n\nANSWER B (Tool 2)\n\"According to Fenwick's 2026 survey of 3,000 US adults, 62% plan to do most holiday shopping online, and Gen Z is the most likely to shop in stores.\"\n\nSURVEY SUMMARY - Fenwick Research, \"2026 Holiday Shopping Outlook\" (Oct 2026)\n- Online survey of 3,000 US adults, Sept 2026.\n- 62% plan to do most of their holiday shopping online (2025: 58%).\n- 41% plan to spend about the same as last year.\n- Gen Z respondents were the most likely to say they'll shop mostly online (71%).\n- 29% say free shipping is the top reason they choose a store.\n\nOur notes: email goes out Nov 20. Legal wants the source named in the email.","rules":[{"type":"numberedItems","count":4},{"type":"maxWords","max":140},{"type":"plainText"}],"mistakes":[{"title":"Not naming the judge","why":"The AI picks the answer that sounds more specific, which isn't the same as correct.","fix":"Add: \"Judge both only against the survey summary.\"","checks":["r0","r1"]},{"title":"Not asking for a usable line","why":"You learn what's wrong but still have to write the stat yourself.","fix":"Add: \"Safe stat: one sentence with the exact number and survey name.\"","checks":["r2"]},{"title":"Long or formatted output","why":"AI writes a side-by-side table with bold.","fix":"Say: \"Exactly 4 numbered lines, each under 25 words, plain text.\"","checks":["numberedItems","maxWords","plainText"]}],"promptHints":[{"label":"Uses the survey as the judge","patterns":["survey summary","only against","judge"]},{"label":"Names the 4 labels","patterns":["a error","b error","safe stat","verdict"]},{"label":"Asks for a paste-ready stat","patterns":["safe stat","paste","exact number"]},{"label":"Asks for the source name","patterns":["survey name","name the source","fenwick"]},{"label":"Asks for plain text","patterns":["plain","no bold","no markdown"]}],"hint":"Two confident answers can both be wrong. Judge them against the survey and ask for one stat you can safely quote.","rubricLabels":["The A error line says A's 72% is wrong (the survey says 62%) and/or that last year was 58%, not 50%","The B error line says B has Gen Z backwards: Gen Z was the most likely to shop mostly online (71%), not in stores","The Safe stat line is one sentence with 62% and the Fenwick 2026 Holiday Shopping Outlook (or Fenwick Research), and the Verdict line says neither answer can be used as is (or names B as closer), coming last"]}