Why does an AI recommend one business over another?
It is not a ranking and it cannot be bought. A model assembles the answer from whatever it can read about businesses in that area …
The shift
How did AI change the way people find local businesses?
The short answer
The question has not changed — people still ask who they should call. What changed is the answer. Instead of a ranked page of ten links to choose between, an assistant now returns two or three names and stops. There is no page two of a conversation, so being on the list is no longer the same thing as being found.
What changedThe answer, not the question
Result formatTwo or three names
Page twoDoes not exist
01 · Why
Why it matters
The question people ask has not changed at all. Everything about how it gets answered has.
The list became an answer
For twenty-five years, finding a local business meant being handed a list and doing the work yourself. Ten blue links, a map pack, some ads at the top. The searcher compared, opened tabs, read a few reviews, and made a decision. Ranking eighth still meant something, because eighth was on the page.
An assistant does not hand back a list. It reads, weighs and decides, then returns a short answer — typically two or three names, sometimes one. The comparison work that used to happen in the customer's head now happens inside a model, before they ever see a result. Ranking eighth in that world is not a worse position. It is absence.
Ranking eighth in that world is not a worse position. It is absence.
Nobody announced it
This did not arrive as an event. There was no morning when search changed and everyone was told. It happened as a gradual change in habit — voice assistants in cars, an AI summary above the results, a chat window that gives a straight answer instead of a page of links — and most business owners have not noticed because their phone still rings.
That is the awkward part of the shift. The decline shows up first as flat enquiries rather than as an obvious drop, because the customers you are losing never appear in your analytics. They asked a question somewhere you cannot see, got somebody else's name, and never arrived to be counted.
02 · What
What it actually is
Three shapes of question, each behaving differently.
Three different kinds of asking
It helps to separate them, because they behave differently.
The recommendation ask — "who's the best plumber near me" — expects curation. It wants one to three names, not a list to work through.
The problem ask — "my AC is making a rattling noise, who do I call" — describes a symptom rather than a service, and the model has to work out which trade solves it before it names anyone.
The verification ask — "is this company any good" — happens after somebody already has a name, from a postcard, a truck, or a friend. It is the one most businesses lose without ever knowing it happened.
03 · How
How it works
Where the answer is actually assembled from — and what that leaves open.
Where the answer actually comes from
A model composing a local recommendation is not consulting a ranking. It is assembling an answer out of whatever it can read and trust about businesses in that area — directory listings, aggregator pages, whatever it retrieved at the time, and any structured data it can query directly.
That is the practical opening. The sources it leans on are mostly third-party pages describing you, written by nobody in particular. A structured record you publish yourself, in a form built to be read by machines rather than skimmed by people, is a source of a different quality — and it is the only one you control.
What this does not mean
It does not mean traditional search is over, and anyone telling you that is selling something. Most local businesses still get most of their customers through Google, a map, or somebody's recommendation.
It means a growing share of the deciding now happens inside a conversation you are not part of, and that share is not going back down. The businesses that will do well are not the ones that abandon what works — they are the ones that make sure there is something of theirs to read when the question gets asked somewhere new.
04 · Questions
Common follow-ups
Is this just SEO with a new name?
No. SEO competes for position on a page of results. This is about whether a model has anything trustworthy to read about you when it composes an answer that has no positions in it. Related work, different mechanism.
How many people actually search this way?
Enough that it shows up in our own probe data across every trade we cover, and growing. But the honest answer is that nobody has a reliable public number, and you should distrust anyone who quotes you one to two decimal places.
Do I need to stop doing what I'm doing now?
No. Nothing here replaces your existing marketing. It covers the channel that existing marketing cannot reach, because ads and rankings do not appear inside an AI answer at all.
Start where AI starts
Try it for 14 days. If it isn't what we said, we refund your first month in full and take the site down.