What Is an AI Visibility Score? How to Read Yours
An AI visibility score estimates how often AI assistants name your business. What feeds the number, how to read it, and why a single run proves little.
Key Takeaways
- 45% of US consumers used AI tools for local business recommendations in the past year, up from 6% (BrightLocal, 1,002 US adults).
- ChatGPT recommended 1.2% of brand locations, against a 35.9% average appearance rate in Google's local 3-Pack (SOCi, 350,000+ locations across 2,751 multi-location brands).
- ChatGPT and Google's AI gave the same brand list on fewer than 1 in 100 repeat runs (SparkToro, 2,961 runs). One test is a sample of one.
- Google says AI Overviews and AI Mode have no additional requirements beyond being indexed and eligible to show with a snippet.
- Our free scan returns four pillar scores from 0 to 100, an overall coverage figure out of 61, and the results of 10 questions put to ChatGPT, Gemini and Perplexity.
What is an AI visibility score?
An AI visibility score is a number that estimates how often AI assistants name or cite your business when someone asks a question you should win. No standard formula exists. Every tool defines its own, so a 40 from one tool and a 40 from another do not mean the same thing.
The number matters because customers now ask. BrightLocal's 2026 Local Consumer Review Survey found 45% of US consumers used AI tools for local recommendations in the past year, up from 6% the year before. The sample was 1,002 US adults across all business categories, so it describes consumers broadly, not roofing or HVAC buyers.
Getting named is hard. SOCi's Local Visibility Index release reports ChatGPT recommended 1.2% of brand locations, against a 35.9% average appearance rate in Google's local 3-Pack. The data covers more than 350,000 locations across 2,751 multi-location brands, and the release does not describe how the AI answers were collected. Use it as a direction, not as your odds.
Why does one test run not tell you your score?
Because AI answers change every time you ask. In SparkToro's study, ChatGPT and Google's AI returned the same brand list on fewer than 1 in 100 repeat runs. A score built from one prompt on one afternoon is a sample of one. Trust a percentage across many runs, never a single result.
That finding comes from SparkToro's study of AI brand recommendations: 600 volunteers ran 12 prompts through ChatGPT, Claude and Google's AI, 2,961 runs in total, in November and December 2025. The authors conclude that visibility percentage across many prompts run multiple times is a reasonable metric, and that a "ranking position" in AI is not.
Read it with the authors' own caveats. The work is not peer reviewed, a co-author works at the AI tracking company Gumshoe.ai, and the prompts covered consumer categories such as headphones, not local service businesses.
What goes into an AI visibility score?
Two kinds of signal, usually blended. Site-readiness checks ask whether AI systems can reach and read your pages. Answer tests ask whether AI names you when it is questioned. The first takes a minute to run. The second needs repeated asking.
| Signal | What it measures | How it is gathered | What it cannot tell you |
|---|---|---|---|
| Site readiness | Crawler access, headings, FAQ content, schema, text readable without JavaScript | One scan of your site | Whether AI actually names you |
| Mention rate | Share of AI answers that name your business | Repeated test questions | Why you were named or skipped |
| Citation rate | Share of answers that link your domain | The same test runs | Whether anyone clicked the link |
| Competitor share | Which businesses get named instead of you | The same test runs | Whether those businesses are better |
How does the Tailwinds Ops scan work out its score?
Our free AI visibility site scan reads your homepage and returns four pillar scores from 0 to 100, plus an overall coverage figure out of 61. Each pillar is the share of its pass-or-fail checks your site passed. A check the scan cannot run is left out, not scored as a zero.
- Technical Foundation (10 checks): HTTPS, http-to-https redirect, valid robots.txt, sitemap, no accidental noindex, canonical tag, mobile viewport, and three Google PageSpeed measures.
- Content and On-Page (7 checks): title length and city, meta description, a single H1, H2 structure, more than 300 words, and image alt text.
- AI Answer Readiness (7 checks): AI search crawlers allowed, a deliberate policy for training bots, question-shaped headings, direct answers under headings, an FAQ block, FAQ or Service schema, and content readable without JavaScript.
- Local Authority (5 checks): LocalBusiness or Organization schema, name, address and phone in the HTML, a tappable phone number, and your city in the title or H1.
After you enter an email and business name, the scan puts 10 questions to ChatGPT, Gemini and Perplexity and records who got named and who got linked. Those calls run through each company's API with live web search on. The results page says plainly that API answers approximate, but do not exactly match, what you would see in the consumer apps, and that answers vary between runs.
How do you read your score?
In two layers. The pillar scores are a fix list. The mention rate tells you whether the fixes are working. A site can pass every check and still go unnamed, so never treat a high pillar score as proof of visibility.
| What you see | What it usually means | Fix first |
|---|---|---|
| AI Answer Readiness low, other pillars fine | AI crawlers blocked, or key text hidden behind JavaScript | Robots.txt and plain-text content |
| Site scores high, named in 0 of 10 answers | Your site is readable, but nothing outside it gives AI a reason to pick you | Reviews, listings and consistent business details off-site |
| Named, but your site is never linked | AI knows the business, and no page answers the question directly | A service-plus-city page with direct answers |
| Named in some answers, missed in others | Normal run-to-run variation | Rerun monthly and track the percentage |
The second row is the pattern to expect once your site is clean. The causes are covered in why ChatGPT doesn't recommend your business.
Which crawler settings affect your AI visibility?
OpenAI runs separate crawlers with independent settings. Per OpenAI's crawler documentation, OAI-SearchBot is for search, and a site that opts out will not be shown in ChatGPT search answers. GPTBot is used to crawl content for training. Blocking GPTBot does not remove you from search. Robots.txt changes can take about 24 hours to take effect.
Google is simpler. Its AI features guidance says there are no additional requirements to appear in AI Overviews or AI Mode. A page must be indexed and eligible to show in Search with a snippet, and no special schema is needed. Our scan checks for schema anyway as a readiness signal. Google states it is not required, and we do not claim schema alone moves your answer-test results.
What should you do with your score this week?
- Run the scan and write down the four pillar scores and the overall figure out of 61.
- Fix any crawler block first. It takes minutes, and it can remove you from ChatGPT search answers entirely.
- Run the manual 10-minute ChatGPT recommendation test to get your own named rate across repeated runs.
- Repeat both on the first of each month and compare percentages, not single answers.
For the market numbers behind this, see our AI search statistics for local businesses. If you would rather have the tracking and fixes handled, that is our answer engine optimization service. Tailwinds Ops is at tailwindsops.com, and the scan is free.
Frequently Asked Questions
What is a good AI visibility score?
There is no standard scale, so a good score depends on the tool. On our scan, a pillar score of 100 means every applicable check passed, which says your site is readable, not that AI names you. Track your named rate across repeated test runs and watch how it moves month to month.
Is an AI visibility score the same as a Google ranking?
No. A Google ranking orders pages for a query. SparkToro's authors argue that a ranking position in AI answers is not a reliable metric and that the share of runs in which you appear is. Their study covered 2,961 runs on consumer categories.
How often should I check my AI visibility?
Monthly, with the same questions each time. Answers change from run to run, so one check tells you little. Log how many runs name you and compare that percentage over time.
Does blocking GPTBot hurt my visibility in ChatGPT?
Per OpenAI's documentation, GPTBot governs training use, and OAI-SearchBot governs search. Each setting is independent. Check your robots.txt for a block on OAI-SearchBot, since that is the one that removes you from ChatGPT search answers.
Do I need special schema or AI files to appear in Google's AI Overviews?
No. Google's guidance says there are no additional requirements, and no special schema.org markup or AI text files are needed. The page must be indexed and eligible to show with a snippet.