As of 13 August 2026, AI can find gaps in your customer FAQs.

What to actually do

1. Hand it to a person

   The route this page recommends

   A person who owns the outcome does this end to end, worth it when the failure is dear.

2. Do it yourself

   Second choice

   A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

   How to actually do it

   1. Open the current FAQ or help-centre pages and copy their text into a document, preserving page titles, links and any dates or version notes.
   2. Export a representative sample of recent support tickets, live-chat transcripts, contact-form subjects and site-search queries, then remove names, email addresses, order numbers and other unnecessary personal data.
   3. Add the current product documentation, policies, price list, eligibility rules and known limitations that staff use when answering customers.
   4. Paste the FAQ content, anonymised customer evidence and supporting documents into a chatbot with the supplied prompt, keeping each source labelled.
   5. Check every proposed gap against the quoted customer evidence and the current source document, then ask a support colleague to confirm priorities and identify questions that need a person.
   6. Update only the approved FAQ pages, compare each drafted answer against the current policy and price list, and record which evidence supported the change.

   Prompt

   Find gaps in the customer FAQs for [COMPANY OR PRODUCT].
   
   I will provide:
   1. The current FAQ or help-centre content.
   2. Anonymised customer support tickets, chat transcripts, search queries or contact reasons.
   3. Any current product information, policies, prices, eligibility rules or known limitations.
   
   Compare the customer questions with the FAQ content. Produce a table with these columns: customer question or topic, evidence from the supplied data, matching FAQ page or answer, gap type, likely customer impact, suggested priority, and recommended next action.
   
   Use these gap types only where they fit: no answer, partial answer, unclear wording, outdated answer, answer is hard to find, conflicting answers, or question should be handled by a person.
   
   Do not invent customer questions, product facts, policies, causes, volumes or priorities. Quote or identify the supplied evidence for every proposed gap. Separate direct evidence from your inference. Mark a finding as uncertain when the supplied data does not support it. Do not rewrite the full FAQs. For each high-priority gap, draft a short proposed answer only when the supplied material contains enough facts, otherwise state exactly what information a colleague must provide. End with the five checks a colleague should complete before publishing any change.

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.

3. Use a tool built for this

   The distant third