Customer Comments Coding Audit: Are Themes Consistent and Evidence-Based?
Check how survey comments have been grouped and counted.
The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.
Audit my coding of customer comments. A sample of 40 comments and the theme I assigned to each: [CODED_COMMENTS]. My list of themes with definitions: [THEME_DEFINITIONS]. How I report results: [REPORTING_METHOD]. Re-code a sample of 15 comments independently and compare with my codes to calculate agreement (matches divided by comments, with the arithmetic). Identify themes that overlap, are too broad, or are used inconsistently, and comments that mention several themes. Make a table with Comment, My code, Second code, Agreement, and Note. Suggest clearer definitions with examples, a rule for multi-theme comments, and a count of themes by frequency with caution about sample size. Recommend how to present results without overstating certainty. Protect customer privacy and do not invent comments. Add a worked agreement calculation: of 15 re-coded comments, 11 match the original code (73 percent), 3 differ because the comment mentions two themes, 1 is ambiguous; propose a rule that the primary theme is the one the customer asked for action on and the secondary theme is recorded in a second column. Show the frequency table of themes and a caution about small numbers.
Fill in before use
Optional: type your details and the prompt above updates. Nothing is saved or sent.
Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.