Short answer: AI has added a way to ask about local businesses, not replaced the map pack. Proximity still decides which firms are candidates; what changed is that distance alone no longer wins you the position, because a nearby competitor with complete, consistent, current details will beat a pin with nothing behind it. The work is mostly unchanged — be findable, be unambiguous, be current.
What actually changed, narrowly
It is worth being precise, because a lot of writing on this treats it as a replacement rather than an addition.
Local search has not stopped working. The map pack, the local results, the route to your door — that is still where a large share of local transactions begin, and it is still distance-aware. Nobody choosing a locksmith at 11pm is reading a generated summary instead of checking whether the business is twenty minutes away.
What is new is the second place someone might ask. And when they do, the answer tends to be a short paragraph assembled from whatever was found — a profile, a page, a few reviews — rather than a map they scroll.
The practical consequence is uncomfortable but simple: your business has to make sense as a paragraph. Name the trade, the area, and what you actually do, in sentences that stand alone. If the only description of you that exists is a map pin and a logo, there is nothing to summarise except what third parties have written about you.
Key takeaway: the change is not that proximity stopped mattering. It is that “we exist” is no longer sufficient — you have to be describable.
The one thing assistants genuinely cannot know
This is the most useful thing about local answers, and it is not a mechanism worth optimising for.
An assistant summarising businesses has no way of knowing whether you can take a job today, whether you are fully booked until Thursday, or whether your van is out. It has whatever was published. Which means a business whose availability and coverage are clearly stated wins against one that leaves the reader to phone and find out — not because it was ranked higher, but because the summary could actually say something useful about it.
Which makes the practical list short:
- State the area you cover, in words, on the page. Vagueness here costs you candidates.
- State hours, and make sure they match your profile.
- State what you do, specifically enough that it distinguishes you from the other three firms in the same category.
- Say what happens when someone gets in touch.
None of that is clever, and all of it is currently under-done.
Consistency beats volume
The single most common local problem is not that a business is invisible. It is that the version of it that a search system finds is wrong or incomplete in some detail.
The same business appears on a website with one phone number, a profile with another, a directory with a third, an old address on a blog post from four years ago, and hours in two places that disagree. Google’s guidance on representing a business is unusually direct about why this matters. [1] It is hard to recommend a business whose details contradict each other, and it is hard to do anything useful about it.
So the first project is not “get more reviews” or “get more mentions”. It is a spreadsheet: every place the business appears, and what each place says. Fix the disagreements first. The mechanics of local visibility are well documented and largely mechanical. [2] Declaring the same facts in markup on your own site — the business type, the address, the hours, the telephone — is what LocalBusiness structured data is for, and it removes the guesswork for anything reading your site rather than your profile. [3] It is unglamorous and it is where most of the available gain sits.
Reviews: what they actually do locally
Reviews do two useful jobs for a local business, and neither involves sentiment analysis.
First, they are how your vocabulary gets written in someone else’s words. If twenty reviews say you turn up the same day and explain the price first, that is a description of you in the words of twenty people who paid you — which is far more useful than anything you could write yourself, because it is not marketing copy.
Second, they are a sign of current operation. A profile with reviews from this month, last month and six months ago is visibly a working business. One with nothing since two years ago is a question mark, and a fair one.
Asking for them is mostly a matter of asking at the right moment, once, plainly, and accepting that not everyone will. [4] Google’s tips on soliciting reviews are worth reading before you write the request. Getting reviews onto your profiles covers the mechanics.
On brand mentions in answers
There is a legitimate practice here, and it is being oversold.
Running a fixed list of questions you expect to be asked, checking the answers, and noticing when you are absent — that is ordinary measurement, and worth doing. Some of the tools for it are described in Ahrefs’ guide to monitoring brand mentions in AI answers. [5]
What is oversold is the idea that mentions can be manufactured. Nobody has a lever here, and a business that has never been mentioned cannot become mentioned through a tactic. What you can influence is whether, when you are mentioned, there is a clear and correct description of you available for whatever is summarising to use — which brings us back to consistency.
What to do first
- Build the spreadsheet of every place your business details appear. Reconcile the disagreements.
- Write one sentence that says what you do and where, and use it more or less verbatim everywhere.
- Check that hours match between your profile and your site. Then set a reminder, because they will drift.
- Ask for reviews after each completed job, steadily, for years rather than in bursts.
- Check whether you are mentioned when you search for yourself. Record the answers monthly and do nothing else clever about it.
How this connects
- A justification web covers attaching checkable facts to claims.
- Clicks versus direct answers covers the broader change this is an example of.
- Automated entities covers keeping your business facts consistent structurally.
- Freshness covers why an old address on a blog post is a live problem.
