How to run an AI visibility audit (not just an SEO audit)

A classic SEO audit checks crawl, index, links, and on-page. An AI visibility audit asks a different question: when buyers ask AI, do you get named?

Step-by-step checklist

1. Prompt set

List 15–30 prompts that map to pipeline (category, alternatives, “best for”, objections).

2. Answer capture

Run each prompt across the engines your buyers use. Record who is cited and who is ignored.

3. Page mapping

For every lost prompt, name the URL that should win — or admit you do not have one.

4. Entity & proof gaps

• Vague hero copy
• Missing comparison pages
• No structured facts (pricing model, integrations, compliance)
• Thin About / authorship

5. Technical blockers

Broken canonicals, noindex accidents, slow or JS-only critical content, messy IA.

6. Punchlist

Severity-rate fixes: launch-blocker content, entity cleanup, then net-new pages.

Publishers are also watching new reporting surfaces (for example Google’s expanding AI search data discussions on marketing blogs like neilpatel.com/blog). Use whatever dashboards you have — then add manual or productised citation checks until the data is trustworthy.

AppTime’s version

Cite is the product loop for steps 1–3. Our services team runs steps 4–6 when you need build capacity.

Request an AI visibility conversation →