AI visibility checklist: does the brand show up in AI answers
19 questions and two prompts to find out in 20 minutes whether ChatGPT, Perplexity, Google AI Mode and Copilot name your company when a client asks who to choose. And what exactly gets in the way if they do not.
The checklist follows the method we use in AI visibility audits: brand gaps, prompts, quotable content, off-site mentions, bot access and measurement.
Check what is already done
19 questions across six steps: from brand gaps to measurement. Every item has a one-minute check that needs no paid tools.
Run the prompts
Two ready prompts for the global market. Three runs in each assistant give a baseline measurement instead of a single random answer.
Read the result
Five typical patterns and what to do in each one. Want the measurement done for you? The audit request is at the bottom of the page.
The checklist at work
A one-time demo: tick what is already done and watch the score and verdict change, then copy a self-check prompt.
19 items in six steps
Check only what you have verified, not what you plan to do. One point per yes.
Where the brand falls behind
Queries and prompts
Content AI can take into an answer
Mentions on third-party sites
Bot access
Measurement
Assistants do not know you or confuse you with others. Start with steps 5 and 6, then with three pages of facts and prices.
- 0-6 pointsYour brand is missing from AI search
- 7-13 pointsBase is there, nothing to quote
- 14-19 pointsRoom to grow in mentions
Ask the assistants who they recommend
The checklist shows what is done on the site. The prompts show the result: who AI names in your niche and what it knows about you. Replace [niche], [city or country] and [brand name] in brackets.
Global market: ChatGPT, Perplexity, Google AI Mode
Run the prompt in English: this is how the same sources your overseas clients read will answer. Start a new chat with web search on, three runs in each system.
I am choosing a [niche, for example a dental clinic or a B2B logistics software vendor] in [city or country]. Name 5 to 7 companies you would recommend, in order of preference. For each company give the official website, one sentence on why it is on the list, and the sources you relied on: pages, reviews or publications. If you are not sure about a company, say so instead of guessing.
What do you know about [brand name] ([website])? Describe what the company does, who it serves, its prices or price range, its strengths and weaknesses, and where you got this information. If you have no reliable data, say that directly.
How to read the result
| What you see | What it means and what to do |
|---|---|
| You are missing from the list, competitors are there | A gap with the competitor. Roundups and reviews describe them, not you. Go to step 4: roundups, reviews, discussions. |
| You are named, but the source is someone else's site | A missed citation. You have no page AI can quote. Step 3: a page with facts, prices and the conclusion up front. |
| AI knows nothing about the brand or confuses it with another company | There are few consistent facts on the web. Start with the same description on the site, in maps and directories (step 4) and check bot access (step 5). |
| AI quotes old prices or services that no longer exist | Pages were not updated, the model works from old copies. Step 3: update and show the date. |
| Answers differ across the three runs | This is normal, source ranking is probabilistic. Count the share: how many answers out of three per system named you. That is the baseline measurement you revisit in a month. |
Write the result down: system, prompt, run number, whether you were named and which source sits next to you. Repeat the same prompts a month after the fixes and compare the mention share. That is AI visibility measurement without paid services.
Quick site readiness check
Six more checkboxes on top of the 19: classic search and AI search at the same time.
Is your site ready for classic and AI search (2026)
FAQ
How long does the checklist take?
About 20 minutes for the checklist itself and another 20-30 minutes for the prompts if you run them three times in three systems. No paid tools needed: a browser, access to robots.txt and GA4.
Why should prompts be run three times?
Source ranking in assistants is probabilistic: the same question in two new chats gives different lists. A single answer means nothing, but the mention share across three runs is a baseline you can compare against in a month.
Why is the prompt in English?
For clients in the US, UAE and Europe assistants answer from English-language sources. Ask in English in a new chat with web search on: this is exactly how your future customers query these systems.
Does the checklist work on any CMS?
Yes. The items about robots.txt, text without JavaScript, update dates and markup check the same way on WordPress, Webflow, Shopify or a custom build. Only the place where the setting lives differs.
What is in a visibility audit and how do I get one?
We run your client questions and three competitors through the major assistants, check bot access and server logs, review the pages AI could quote, set up measurement and give a 30-day plan. The first look is free: send a request below.
No time to check it all? We run the measurement
The same checklist and prompts, but with your real client questions, three competitors and a 30-day plan of fixes.
- Your 10 client questions in ChatGPT, Perplexity and Google AI Mode, three runs each: who is named, with which sources, are you there
- Three competitors on the same questions: where they are quoted and you are not, and which pages are responsible
- Bot access: robots.txt, CDN and firewall rules, text without JavaScript, server logs for 7 days
- Review of the pages AI could quote and what they lack: answer up front, prices, dates, markup
- Measurement setup: GA4 segment of AI referrals and an "AI assistant" option in the lead form
- A prioritized 30-day plan and a re-measurement in a month with the same prompts