Skip to main content
When someone asks ChatGPT, Claude or Perplexity for “the best [category] in [your city]”, a handful of businesses get named. The Are you found by AI? framework starts with the answer a buyer actually receives. Does it name your business, and who else does it name?

​
One percentage, with the counts beside it

The headline measure is the named share: countable saved answers that name the business, divided by all countable saved answers. If AI names a business in 7 of 28 countable answers, its named share is 25%.
  • Every saved repetition counts. An engine may be asked the same question more than once, and each saved answer counts on its own. There is no majority vote.
  • A link to the website on its own does not count as the business being named in the answer.
  • Missing or incomplete evidence is left out, never counted as a no.
Read the saved answers alongside the number. The questions, engines, date and answer count tell you what was measured. A short free scan is a smaller sample than a full check.

​
Site checks explain what to work on

Crawler access, readable content, business facts, structured data and independent mentions are separate checks. They are not extra percentages of AI recommendations, and an older weighted score cannot be converted into a named share. The report groups the on-site checks into AI Readiness (out of 100), measured from your site alone and composed of five site signals:
How quotable and self-contained your content is, and how easily an engine can lift a clear, factual passage into its answer. Answer real buyer questions directly, lead with the conclusion and back claims with specifics.
Whether your site shows who is behind the business and why customers can trust it: named people and credentials, a real About page, visible contact details and provenance.
Whether AI crawlers can cleanly read and load your website. Serve content in the initial HTML response: most AI crawlers fetch raw HTML and do not run JavaScript.
Structured data (JSON-LD) such as Organization and LocalBusiness that states what you are, where you operate and what you offer. Deliver it in the server HTML or inject it at the edge, not with client-side JavaScript.
Your published crawler permission in robots.txt, a published llms.txt and answer-shaped content.
A signal that could not be verified (the page could not be read, the site is a JavaScript shell, or robots.txt was unreadable) is left out of Readiness rather than scored zero, so which signals count can differ from scan to scan. Each report states the share each signal carries. Your off-site presence is scored separately as Footprint (out of 100): Wikidata and Knowledge Graph entity, Google Business Profile, brand mentions, search demand and referring domains. Footprint is scored beside Readiness, not inside it. See AI Readiness and Footprint. Reports and the API also carry an older AI Visibility score out of 100. It is a weighted diagnostic score, not a percentage of AI recommendations, and it cannot be converted into a named share. See AI Visibility vs AI Readiness.

​
What to check next

1

Your business facts

Make your name, services, locations and contact details clear and consistent.
2

Your buyers' questions

Give useful answers on pages people can read and share.
3

Independent mentions

Review the sources saved with the answers and identify relevant places where buyers could find you.
4

Access and structure

Check what a crawler receives, the site’s published access rules and its structured data.
5

The next measurement

Compare the same questions and engines over time. Keep changes in coverage separate from changes in results.
These checks guide the work. None guarantees a recommendation, and a change in an answer does not by itself prove which edit caused it.

​
See measured business results

The public benchmarks count saved answers directly. Each business gets a percentage: answers that name it divided by countable answers. Missing evidence is left out, never treated as a negative answer. The current public snapshot was compiled on 29 September 2026. It covers 271 businesses in Australia, 134 in the United States and 105 in the United Kingdom. Different businesses have different questions, dates and answer counts, so these are observed results, not a controlled comparison. See Benchmarks.
Earlier scoring research. Earlier articles compared weighted visibility and site-check scores. Those values were not percentages of answers and are not carried into the current benchmarks. Current pages use counted answers or state that no comparable count is available.

​
Open method

The readiness method and the audit suite are open source at github.com/techhorizonlabs/thl-open. The measurement engine, question sampling and regression eval set are not in that repository. The term GEO comes from Aggarwal et al., “GEO: Generative Engine Optimization,” ACM SIGKDD 2024.