{"id":"cx-refund-decision-class-cancellation","audience":"customer-support","category":"support-work","technique":"verification","track":"support","level":"intermediate","title":"Compute a class cancellation refund from a tiered policy","free":false,"task":"Write exactly three lines: \"Notice:\", \"Refund:\" and \"Reply:\". Notice counts whole days between the cancellation request and the class start. Refund shows the percent from the policy tiers and the dollar math, based on what the customer actually paid. Reply is one sentence for the customer that starts with \"Hi Noor,\" and has no apology words. Plain text, under 60 words.","materialLabel":"Your material (attached to your prompt automatically)","material":"CLAYWORKS STUDIO CANCELLATION POLICY\n- 7 or more days before the class: full refund\n- 3 to 6 days before the class: 50% refund\n- Less than 3 days before: no refund, studio credit for the same amount instead\nRefunds go back to the original payment. Refunds are based on the amount paid after any discount.\n\nBOOKING #2271: Wheel Throwing Basics, Friday Nov 21, 7:00 PM. List price $80. Customer used promo SAVE25 (25% off), paid $60.\nRequest received: Sunday Nov 16, 3:00 PM by Noor Alvi: \"Please cancel and refund my class.\"\n\nNote: count calendar days from the request date to the class date.","rules":[{"type":"plainText"},{"type":"maxWords","max":60}],"mistakes":[{"title":"Using the list price","why":"The AI refunds half of $80 instead of the $60 she paid.","fix":"Say: \"Base it on the amount paid after discount.\"","checks":["r1"]},{"title":"Wrong tier","why":"The AI counts from the wrong date and picks full refund.","fix":"Say: \"Notice = days from the request date to the class date.\"","checks":["r0"]},{"title":"Apologizing in the reply","why":"The AI adds 'we're sorry' even though the policy is clear.","fix":"Add: \"No apology words.\"","checks":["r2"]},{"title":"Getting markdown back","why":"AI adds bold and headings by default, which show up as stray symbols where this is 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 well over the limit.","fix":"Cap each part instead of the total, for example \"each line under 15 words\".","checks":["maxWords"]}],"promptHints":[{"label":"Asks for plain text","patterns":["plain","no bold","no markdown","no formatting","no headings"]},{"label":"Sets a length limit","patterns":["\\d+\\s*words","word (limit|count)","under \\d+"]},{"label":"Names the labels","patterns":["notice:","refund:","reply:"]},{"label":"Uses the amount paid","patterns":["amount paid","paid","after discount"]}],"hint":"Tiered policies need the right date count and the right base amount. Name both and ask for the math.","rubricLabels":["The output is exactly three lines starting Notice:, Refund: and Reply:, and the Notice line says 5 days (Nov 16 to Nov 21)","The Refund line shows 50% of $60 = $30, based on the amount paid, not the $80 list price","The Reply line is one sentence starting with \"Hi Noor,\" that says $30 will go back to her original payment and contains no apology words such as sorry or apologize"]}