{"id":"cs-conflicting-sources-incident-timeline-table","audience":"customer-success","category":"support-work","technique":"verification","track":"support","level":"advanced","title":"Rebuild an incident timeline when the sources disagree","free":false,"task":"Build a plain-text table with the header row exactly \"Date | Event | Source used | Conflict?\" and one row per event in date order. When the ticket system and someone's notes disagree, use the system date and mark Conflict? as Yes; otherwise No. Plain text, under 110 words, nothing outside the table.","materialLabel":"Your material (attached to your prompt automatically)","material":"SOURCES FOR THE KESTREL SYNC INCIDENT\n\nTICKET SYSTEM (timestamps)\n- Oct 21: ticket opened by the customer.\n- Oct 23: priority raised to P1.\n- Oct 29: fix deployed and ticket resolved.\n\nCSM NOTES (written from memory on Nov 3)\n- \"Priority went to P1 on Oct 24.\"\n- \"Fix went out Oct 28.\"\n- \"I offered a service credit on a call on Nov 3.\" (no record elsewhere)\n- \"Customer opened the ticket on Oct 21.\"\n\nCUSTOMER EMAIL (Oct 30)\n- \"Reports are failing again as of today.\"\n\nRule: system timestamps beat memory. A customer's email is its own source.","rules":[{"type":"plainText"},{"type":"maxWords","max":110}],"mistakes":[{"title":"Using the notes' dates","why":"The AI trusts the CSM's memory for P1 and the fix.","fix":"Say: \"Use the system date when sources disagree.\"","checks":["r1"]},{"title":"Marking everything a conflict","why":"The AI flags rows that only one source mentions.","fix":"Say: \"Conflict is Yes only when sources give different dates.\"","checks":["r2"]},{"title":"Missing the customer email","why":"The AI drops the Oct 30 failure.","fix":"Say: \"Include the customer email event with its own source.\"","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":"The AI runs long and cannot count its own words.","fix":"Cap each part (\"each line under 12 words\") and add \"under N words in total\".","checks":["maxWords"]},{"title":"Leaving a requirement unstated","why":"The AI follows its own defaults unless the prompt spells this out.","fix":"State it plainly in your prompt: the header row is exactly 'Date | Event | Source used | Conflict?' with five rows in date order and nothing outside the table.","checks":["r0"]}],"promptHints":[{"label":"Gives the exact header","patterns":["date \\|","header"]},{"label":"System beats memory","patterns":["system","timestamp","memory"]},{"label":"Defines the conflict flag","patterns":["conflict","yes","disagree"]},{"label":"Orders by date","patterns":["date order","chronolog","sort"]},{"label":"Asks for plain text","patterns":["plain","no bold","no markdown","no formatting","no headings"]},{"label":"Sets a word cap","patterns":["\\d+\\s*words","under \\d+","word"]}],"hint":"Say which source wins and exactly when a row counts as a conflict.","mission":"Rebuild the incident timeline where ticket and notes disagree.","musts":["System date wins","Mark conflicts Yes","Under 110 words"],"rubricLabels":["The header row is exactly \"Date | Event | Source used | Conflict?\" with five rows in date order and nothing outside the table","Oct 23 (P1 raised) and Oct 29 (fix deployed) use the ticket system dates and are marked Yes, not Oct 24 or Oct 28","Oct 21 (ticket opened), Oct 30 (customer reports failures, source customer email) and Nov 3 (credit offered, source CSM notes) are marked No"]}