The symptom is specific: aI interview summaries invent customer quotations. Start with the affected item and identify the decision or input that could produce this behavior.
Find the likely cause
The summary rewrites paraphrases as direct speech.
Treat this as an explanation to verify against the actual work. Look at the input, the relevant decision, and the final result together. If the evidence does not support this diagnosis, investigate the mismatch before applying a convenient but unrelated fix.
Make the targeted correction
Keep verbatim quotes in a separate field with transcript locations.
Interview analysis becomes useful when quotes, interpretations, and proposed actions remain distinguishable.
Check that the repair worked
Check every published quote against the original recording or transcript.
Repeat the check on the final version that the reader or customer will encounter. An approved draft, a preview, and a published result can differ; the acceptance decision should concern the version people actually use.
Prevent the next related failure
A separate issue to watch for is this: leading questions distort customer insights. Replace approval questions with requests to describe recent behavior.
Monitor the useful outcome
Measure the share of messaging decisions supported by traceable quotes, and review whether new inquiries fit the intended audience.
Keep a short record of the original symptom, the evidence behind the diagnosis, and the result of the acceptance check. That record makes the solution reusable when the same condition appears again, without assuming that every superficially similar problem has the same cause.