Knowledge Base Answer Accuracy Audit Against Tickets
Check help articles against what agents say.
The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.
Audit knowledge base accuracy. Five help articles and ten recent tickets on the same topics with agent answers: [ARTICLES_TICKETS]. Product changes in the last quarter: [PRODUCT_CHANGES]. Last review dates: [REVIEW_DATES]. Compare article statements with current product behaviour and agent answers. Flag outdated screenshots, steps that no longer work, contradictory advice, missing cases and articles agents avoid. Return a table with Article, Problem, Severity (red, amber, green) and Fix, an owner and review rhythm, and a short update template for agents who find errors. Also rank the articles by risk if wrong, such as billing or safety, and review those first. Suggest a rule that any product release triggers a review of linked articles within a week. Also suggest tagging each article with the product area and the person responsible, and a reminder in the release checklist to review affected articles before launch. Add a note on how agents can flag an error quickly without leaving the ticket. End by suggesting a quick way to explain the findings to a senior person who has two minutes, with the three points to lead with.
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.