AI Visibility Audit: A Checklist Owners Can Run This Week

An AI visibility audit is a check of whether ChatGPT, Perplexity, Gemini, Claude and Google AI Overviews name your business when customers ask for what you sell, and whether what they say about you is true.

Key points

• Start with access. If the engines' search crawlers can't read your site, the rest of the audit matters much less.
• Measure how often you're named across repeated runs of the same question, not where you "rank" in one answer.
• Check accuracy as hard as presence. A wrong phone number or a service you dropped years ago costs you calls.

What is an AI visibility audit?

An AI visibility audit is a structured test of how answer engines see your business. You ask ChatGPT, Perplexity, Gemini, Claude and Google AI Overviews the questions your customers ask, record whether you're named, what's said about you and which sources are cited, then trace each gap back to a page, listing or crawler setting you can fix.

Most audit guides online are written for software brands. This one is for a dental office, a law firm, an HVAC company or an accounting practice that serves a town or a region.

Why run an AI visibility audit now?

Because customers are already asking. BrightLocal's Local Consumer Review Survey, published in February 2026, found that use of ChatGPT and other generative AI tools for local recommendations rose from 6% to 45% in a year. The same survey found that 82% of consumers read AI-generated review summaries.

You should know what the engines tell them. Our guide on how to get recommended by AI when local customers ask around covers more of that data.

How do you run an AI visibility audit yourself?

1. Check that the engines' crawlers can read your site.
2. Write a list of the questions your customers really ask.
3. Run each question more than once in each engine.
4. Record who gets named, what's said about you and which sources appear.
5. Trace every gap to a specific fix.

Step 1: Check that engines can read your site

OpenAI's documentation is direct about this: "Sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers." See OpenAI's overview of its crawlers.

Ask whoever runs your website to confirm that robots.txt and any firewall setting allow OAI-SearchBot (ChatGPT), PerplexityBot (Perplexity), Claude-SearchBot (Claude) and Googlebot.

For Google, check that your main service pages are indexed. Google's page on AI features and your website says a page must be indexed and eligible to show with a snippet to appear as a supporting link in AI Overviews or AI Mode. It also says you don't need special AI text files, markup or schema to appear there.

Step 2: Write the questions your customers ask

Pull the wording from real calls, intake forms and emails, not from your marketing copy. For a local or professional-services business, cover four kinds of question:

• Category and place: "Who's a good estate planning attorney in Dayton?"
• Problem: "My furnace is making a banging noise. Who should I call in Boise?"
• Comparison: "Is [your business] or [competitor] better for a small kitchen remodel?"
• Name check: "What do you know about [your business] in [your town]?"

The first three test whether you're recommended. The name check tests whether the engine has your facts right.

Step 3: Run each question more than once

A single screenshot tells you very little. In research published on January 28, 2026, SparkToro had 600 volunteers run prompts 2,961 times across ChatGPT, Claude and Google's AI. It found less than a 1 in 100 chance that ChatGPT or Google's AI would return the same list of brands in any two responses. The researchers concluded that the share of runs where a brand appears is a reasonable measure, and that ranking positions aren't.

So run each question several times per engine, in fresh chats, and count how often you're named.

Hypothetical example: you run "best family dentist in [your town]" ten times in ChatGPT and your practice is named in four of the answers. Your mention rate for that question in that engine is 40%. That number is worth tracking month to month. "We were third on Tuesday" isn't.

Step 4: Record what you see

Keep one row per run in a simple spreadsheet:

The "sources shown" column tells you which pages and directories the engine leaned on.

Step 5: Trace every gap to a fix

Each pattern in your sheet points somewhere:

• Never named, even in the name check: go back to Step 1. The engine may not be able to read you.
• Named in the name check but not for category questions: your pages probably don't say plainly what you do and where. Give each main service its own page and list the towns you serve in text. Our piece on how to rank in ChatGPT when customers ask who to hire goes deeper on this.
• Named, with wrong facts: find where the error lives. It's usually an old directory listing, an outdated page or a stale Google Business Profile. Google's guidance tells businesses to keep Business Profile information up to date.
• Competitors named, citing sites you're missing from: claim and complete your listing on those directories and review sites.

What should an owner fix first?

Fix access first, because a blocked crawler hides every other improvement. Then fix wrong facts, which do damage even when you're named. Clearer service pages and fuller outside listings come last, since they take longer to pay off.

Can you see AI traffic in Google Search Console?

Not separately. Google says traffic from its AI features is included in your overall search traffic in Search Console, reported in the Performance report under the "Web" search type. It isn't broken out on its own.

So the question list from Step 2 is your real measuring stick. Rerun it on a schedule and compare.

Should you do it yourself or hire an AI visibility audit?

Every step above can be done without paying anyone. The limit is volume: a handful of questions shows the big problems, not a reliable mention rate across many questions, engines and competitors.

If you want that wider view, Vector Index's AI visibility audit measures how your business appears across the engines and gives you a prioritized list of fixes. You or your web person make the changes.

FAQ

How often should I run an AI visibility audit?

Monthly is a sensible rhythm for most owners, plus once after any big website change, move or rebrand. Use the same question list each time so the results are comparable.

Why does ChatGPT give a different answer every time I ask?

Recommendation lists vary from run to run, as SparkToro's research showed. That's why you count how often you're named rather than trusting one answer.

Do I need an llms.txt file or special schema to show up?

For Google AI Overviews and AI Mode, Google says no: no special AI text files, markup or schema are required. For every engine, crawler access is the step to check first.

What's the difference between an AI visibility audit and an SEO audit?

An SEO audit checks whether your pages can be crawled, indexed and ranked. An AI visibility audit checks what the engines say about you, whether it's accurate, and which sources shaped it.

An AI visibility audit doesn't need special tools to begin: confirm the engines can read your site, ask the questions your customers ask, count how often you're named, and fix the facts they get wrong.