Learn · Guide
Why AI assistants know your name but do not recommend you
Ask an AI assistant about your business by name and it will often describe you correctly. Ask it the question a buyer actually types — the best option near them, the one with transparent prices, the one that is open today — and your name is missing. These are two different tests, and passing the first says little about the second.
Updated 19 October 2026
Two questions, two different tests
A question that contains your name answers itself: the engine searches for you, finds your site or a directory listing, and summarises it. That measures recognition. A buyer's question does not contain your name. The engine has to decide who belongs in the answer, from whatever it read for that question. That measures recommendation, and it is the only one that brings customers.
This is why a NamedFirst measurement keeps the two apart. Brand-name questions are asked separately and excluded from the score, because including them would make every business look visible. The score is built only from questions a buyer would type without knowing you exist.
Why recognition does not turn into recommendation
When an engine recommends a business, it justifies the choice with something it can quote. In the answers we measure, that is most often a rating and a number of reviews on a platform, a stated price, a credential, an award, or a direct match between a page and the question. Being known supplies none of these. A business can be well established and still give the engine nothing to point at.
- The evidence lives on sites the engine did not read for that question, so it cannot be used.
- The evidence exists only on your own site, and the engine preferred independent sources.
- A competitor has the same kind of evidence in larger amounts, such as more reviews on the platform the engine read.
- Your pages describe the business but do not answer the specific question, so they are read and not used.
How to see which one is happening
Every engine answer we store comes with the sources it opened or cited. Comparing those sources with the businesses it named shows the pattern directly: which websites the engine trusts for your category, whether your own site was read at all, and what reasons it gave for the names it chose.
| What you see | What it usually means | Where to start |
|---|---|---|
| Your site is never read | No page matches the question, or crawlers are blocked | A page that answers the question; check robots.txt |
| Your site is read but you are not named | The page does not answer directly or lacks quotable facts | An answer in the opening lines, with prices, hours and credentials |
| Competitors are named for their reviews | The engine relies on a review or booking platform | An active profile and recent reviews on that platform |
| Directories describe you differently | Conflicting facts across sources | Make name, address, services and prices consistent |
What closes the gap
The work is evidence, not existence. Start with the sources the engines already read for your questions, because that is where a recommendation is decided. Then make your own pages answer those questions in a form an engine can lift: a direct opening answer, a price, a list, an opening time, a named accreditation. Finally, keep facts consistent everywhere, so the engine has no conflict to resolve against you.
Questions
Common questions
Keep reading