AI visibility
How firms get into the answer.
Nearly half of U.S. consumers used AI to find a local business last year, and the answer names about three. These guides explain what decides which ones, engine by engine, with every figure sourced. Written for the owner or partner who has typed their own specialty into ChatGPT and not seen their name.
For what changed this week, see the daily AI Industry Updates.
What is AI visibility?
AI visibility is how often an AI engine names your firm when someone asks it for a recommendation in your specialty and market.
Search engines rank pages. AI engines compose a short list, about three firms at the median in the answers we measure, and describe each one. Being outside that shortlist is not a lower position, it is absence. So AI visibility is measured as presence and prominence inside an answer rather than as a rank in a list.
The shift is recent and steep. 45% of U.S. consumers used AI tools like ChatGPT to find a local business in the past year. That number was only 6% the year before (BrightLocal, Local Consumer Review Survey 2026), and 31% now do it at least once a month (BrightLocal, July 2026). The machines are choosy: 1.2% of 350,000+ business locations got a ChatGPT recommendation, while Google's top three map results showed 35.9% of them (SOCi, 2026 Local Visibility Index). Showing up on Google no longer means being named.
Is AI visibility the same as SEO?
No, though it depends on some of the same groundwork. SEO improves where a page of yours sits in a list of links. AI visibility depends heavily on sources you do not own. Muck Rack analyzed 25 million+ links that ChatGPT, Claude and Gemini pointed to: 84% led to earned media, sources the business didn't write or pay for, and paid placements made up 0.3% (Muck Rack, What Is AI Reading?, 2026).
Ahrefs studied 75,000 brands and found the strongest link to AI visibility was how often a brand was mentioned on YouTube (about 0.737, where 1.0 is a perfect match), while backlinks showed only very weak links (Ahrefs, December 2025). A crawlable site remains a precondition: a review of 54 studies scored being crawlable 9.5 out of 10 (Zyppy, May 2026), and 73% of sites have technical barriers that block AI crawlers (Otterly, 2026). But a technically perfect site can still be absent from every answer.
Which AI engines matter, and do they agree?
Five matter for professional firms, and they disagree with each other enough that visibility has to be measured separately in each.
They also trust different sources. Brand websites made up 59.8% of Google AI Overviews' sources and 28.9% of Perplexity's (Otterly, 2026). A firm that shows up in one can be missing from the next, which is why a single screenshot of a good ChatGPT answer tells you almost nothing about your position.
What decides whether your firm gets named?
Read the firms that do get named for a given specialty and the same four things appear every time.
A clear entity
One name, one address, one phone number, one specialty, consistent across your site, your Google Business Profile, your directories and your professional listings. This is how a machine confirms you exist.
Third-party proof in the sources your engine trusts
Every profession has a short list the engines already treat as authoritative, and it differs by field: legal directories for a law firm, the certification lookup and specialty society for a surgeon, SART for a fertility clinic.
Published expertise with the receipts attached
Adding statistics, quotations or named sources raised a page's visibility in AI answers by 30 to 40% in a peer-reviewed study (Aggarwal et al., KDD 2024). A review of 54 studies scored "factually specific" 8.3 out of 10 and "cites sources" 8 (Zyppy, 2026).
A reputation that passes
Locations that ChatGPT recommended averaged 4.3 stars (SOCi, 2026). You don't need the most reviews in your market. You need a rating that holds up.
The full version of this, with every figure sourced and the per-profession directory lists, is in Why your firm isn't recommended by AI.
Can anyone guarantee an AI will recommend you?
No, and it is worth being blunt about it. AI answers vary between engines, between users and between runs, and no provider controls what a model outputs. Anyone promising a position in the answer is promising something they cannot deliver, and in most licensed professions it is a claim your governing body would not let you repeat. Google's own guide to AI features, updated July 2026, makes a related point about method: it lists llms.txt files, "chunking" and special AI-friendly formats among the things you do not need, and names people-first expert content, a crawlable site and a maintained Google Business Profile among the things that matter.
How do you find out where you stand?
You measure it the way a client would. Take the real questions people ask for your specialty in your market, run them across all five engines, and record the answers word for word: who was named, in what order, and which sources were cited. Score that and freeze it as a baseline, so every later claim can be checked against something that cannot move. Our full method, including the weights and the prompt-set design, is published at how we measure.
The free Visibility Audit does exactly that and returns a 0 to 100 score, the actual AI answers for your specialty and market, and your three largest gaps, in a 30-minute readout. One audit per firm. Before you request one, worth knowing: we take one firm per specialty, per market, and you can check whether your field is still open in about ten seconds at forevercited.ai/availability.
The guides
Longer answers, fully sourced.
The free audit
Find out what AI says when your next client asks.
Your 0–100 Visibility Score, the actual AI answers for your specialty and market, and your three biggest gaps, free, in a 30-minute readout. One audit per firm.