When buyers ask AI who to choose, does it name you?
Legible puts the questions a buyer would ask to 7 AI search tools, counts how often each one recommends you and who it recommends instead, and shows what to change. Nothing goes live until a person approves it.
“best compact tractor dealer near Des Moines”
One dot is one AI answer: 13 buyer questions across, 7 AI tools down. A fictional dealer, to show how a report reads.
How it works
Four steps, repeated. The same questions every time, so one check can be compared with the next.
Ask
Legible writes 13 buyer questions from your site, and you can change them. They go to AI Overviews, AI Mode, ChatGPT, Gemini, Perplexity, Claude and Grok. They do not name you, so each answer shows whether AI brings you up on its own.
Read
Every answer is read for who it recommends and in what order, which websites it cites, and what it gets wrong about you.
Fix
You get a short list of what to do first: wrong facts to correct, the lists and articles AI reads, and ready-made changes to your site. A person approves each change before it goes live.
Check again
The same questions run again, so you see a trend instead of one day's answers, along with a likely range around every rate.
A report a business owner can read
Four screens: Score, Why, Fixes and Progress. Open the example to go through them, or read the same report as a PDF.
What we will not tell you
This field is full of promises. These are the limits we state up front, and the report states them too.
- We do not promise a ranking.AI answers change from one day to the next, and your rivals move too. Every rate comes with a likely range.
- A readiness score is not visibility.Legible checks how easy your site is for AI to read, and it reports that separately. A site can pass every check and still not be named.
- These numbers depend on more than your site.In Ahrefs' study of 75,000 brands, mentions on other websites went with appearing in Google's AI Overviews. That is a correlation, not a guarantee. Legible shows the websites AI cites and what to do on each.
- It measures the brand, not each location.A site gets one set of questions for one market, so a business with several locations is measured as a whole. Separate countries can be measured apart, and a single city can be set, but not one result per branch.
- One check is not a trend.A change is only called up or down at a pilot's planned end, and only if it goes beyond what your rivals did on the same questions. Until then the report says it is too early.
Start with a 12-week pilot
The pilot locks a larger set of questions before anything changes, then checks every week. At the planned end you get one verdict: up, down, or inconclusive. It takes a clear lift across many questions to call a change, and smaller shifts read inconclusive. That is a real result too, and we will say so.
Schedule a demo- buyer questions
- 50
- rivals compared on the same questions
- Up to 5
- weekly checks before the fixes
- 3
- weekly checks after 3 weeks for answers to settle
- 6
Questions people ask
What is AI visibility?
How often AI search tools name or recommend a business when buyers ask about its category. Legible measures it by asking 7 tools a fixed set of questions that do not name the business.
Which AI tools does Legible check?
AI Overviews, AI Mode, ChatGPT, Gemini, Perplexity, Claude and Grok. Answers come through each tool's developer access with web search available, so they can differ from what you see in the app on your phone.
Will Legible change my website without asking?
No. Every fix waits for a person to approve it. Publishing is refused until those approvals exist, and changes outside your site, such as listings and outreach, are made by a person.
How does a fix reach my site?
Through a WordPress plugin, a GitHub pull request, or steps your web team applies by hand. Legible then checks the live site to confirm each fix is being served.
How soon will the numbers move?
Usually not in days. AI answers vary from run to run, so a real change takes weeks of checks to show. That is why the pilot runs for 12 weeks, and its result can be inconclusive.
Is this the same as SEO?
They overlap, but the question is different. SEO asks where a page ranks. Legible asks whether AI names you in its answer, and which websites it read to get there.