{"id":"res-cs-csat-final-huddle","audience":"customer-support","category":"research","technique":"audience","title":"Share the final CSAT results with your agents in three sentences","free":false,"task":"You're opening Thursday's agent huddle with last month's satisfaction results, and you're running a refresher on the new refund macro right after. You want at most 3 sentences in plain text that give the final CSAT score, the most common reason for low ratings with its count, and mention the refund macro refresher. Use only the final report, not the mid-month snapshot, and don't name individual agents.","materialLabel":"The CSAT report and a mid-month snapshot (attached to your prompt automatically)","material":"MID-MONTH SNAPSHOT (Oct 15, from the dashboard, partial month): CSAT 81% from 540 surveys. Posted in the team channel: \"Looking rough, team!\"\n\n-----\nFINAL CSAT REPORT - October 2026 (Support Ops)\nSurveys returned: 1,260\nCSAT (satisfied or very satisfied): 88%\nLow ratings: 151\nTop reasons for low ratings (from comments):\n- Refund took too long to explain or process: 64\n- Had to repeat information to a second agent: 41\n- Wait time over 10 minutes: 29\n- Other: 17\nAgent with the most low ratings: Jordan P. (12)\nAverage first response time: 6 minutes (September: 9 minutes)\nNote: CSAT counts customers who chose 4 or 5 on a 5-point scale. Individual agent results are shared privately by team leads, not in group settings.","rules":[{"type":"maxSentences","max":3},{"type":"plainText"}],"mistakes":[{"title":"Using the snapshot","why":"The mid-month 81% comes first and sounds dramatic.","fix":"Say: \"Use only the final October report.\"","checks":["r0"]},{"title":"Calling out an agent","why":"The report names one agent, and the AI may include it as a 'key finding'.","fix":"Add: \"Don't name individual agents.\"","checks":["r2"]},{"title":"Missing the link to training","why":"The AI doesn't know about your macro refresher unless you mention it.","fix":"Add: \"Mention the refund macro refresher.\"","checks":["r1"]},{"title":"Not capping the length","why":"The AI writes several paragraphs by default, far more than the reader will take in.","fix":"Say: \"at most 3 sentences.\"","checks":["maxSentences"]},{"title":"Getting markdown back","why":"AI defaults to bold labels and headings, which show up as stray symbols in your huddle script.","fix":"Ask for \"plain text, no bold, no headings.\"","checks":["plainText"]}],"promptHints":[{"label":"Uses the final report","patterns":["final","not the (snapshot|mid)","88"]},{"label":"Protects individuals","patterns":["(don't|do not) name","no names","individual"]},{"label":"Mentions the refresher","patterns":["refresher","macro","training"]},{"label":"Caps the number of sentences","patterns":["\\d+\\s*sentences","(two|three|four|five) sentences","sentence"]},{"label":"Asks for plain text","patterns":["plain","no bold","no markdown","no formatting","no headings"]}],"hint":"Team results land better when they're accurate and nobody gets singled out. Tell the AI which report is final, what to leave out, and what comes next.","rubricLabels":["Gives the final CSAT of 88% (not the mid-month 81%)","Names refunds as the top reason for low ratings with its count (64), and mentions the refund macro refresher","Does not name any individual agent (no mention of Jordan P.), and has no headings, greeting or sign-off"]}