AI Search Visibility: How to Measure Whether ChatGPT, Gemini and Perplexity Recommend You
Quick answer
To measure AI search visibility, build a fixed set of the questions your buyers ask, run them in ChatGPT, Gemini and Perplexity every month, and record whether you are named, whether the description is correct and which sources are cited. The share of answers that name you is your visibility score. Track it per assistant and fix the gaps the citations reveal.
On this page
Why measure at all
Buyers now ask assistants for shortlists: "best dental implant clinic in Istanbul", "who makes hydrocarbon dew point analysers". If you are not named, you are not considered, and you will never see that lost enquiry in your analytics. Measurement turns an invisible problem into a number you can move.
Step 1: build a prompt set
- Collect 20 to 50 questions from sales calls, emails and search terms.
- Mix the stages: problem questions, comparison questions and "who should I hire" questions.
- Write them the way buyers speak, including location where it matters.
- Freeze the set. Changing prompts every month makes trends meaningless.
Step 2: run and record
Run each prompt in each assistant in a clean session. For every answer, record four things: were you named, were you recommended, was the description accurate, and which sources were cited. Assistant answers vary between sessions, so run important prompts more than once and record the share.
- ChatGPT75%
- Perplexity95%
- Gemini35%
Share of tracked prompts where the brand appeared. Source: Zegaz AI visibility tracking, 2026.
The example shows why you track per assistant. The same company can be strong in one and weak in another, because each assistant leans on different sources.
Step 3: read the citations
Citations tell you what each assistant trusts. If competitors are cited from an industry directory, a trade publication or a comparison article, that is where you need a presence. If your own pages are cited but the description is wrong, the fix is on your site.
Note that paying for ads does not buy citations. A June 2026 SE Ranking study found only 11.53% of advertiser domains in Google AI Mode's cited sources, with no advantage over sites that did not advertise.
Step 4: turn gaps into fixes
- Pages that answer each prompt directly, near the top.
- Organisation, Service and FAQ schema.
- Identical business facts across your site and profiles.
- Mentions on the sources the assistants cite.
- Fresh proof: case studies with dates and numbers.
Then rerun the set next month and compare.
What is a good score?
There is no universal benchmark yet, because every prompt set is different. Compare yourself with your top three competitors on the same prompts, and with your own score last month. Direction matters more than the absolute number.
How OUNTERNET helps
Our free AI visibility check runs a short prompt set for your business. The full AEO and GEO programme tracks a larger set monthly and fixes the gaps, from £2,900 a month, with a one-off AI Visibility Audit from £2,500. Prices exclude VAT. Final price after we analyse your business and scope.
Frequently asked questions
What is AI search visibility?
How often AI assistants such as ChatGPT, Gemini and Perplexity name, recommend and cite your business when buyers ask relevant questions.
How many prompts should I track?
Twenty to fifty is enough for most businesses, as long as the set stays fixed so trends are comparable.
Why do results differ between assistants?
Each assistant uses different sources and retrieval methods, so the same business can score very differently in each.
Can I pay to be cited?
No. Ads are separate from answers, and studies show advertisers get no citation advantage.
How often should I measure?
Monthly works for most businesses. Run key prompts more than once because answers vary between sessions.