Growth
Your next customer
will ask an AI
about you.
Search stopped being a list of links and became an answer. For a local business that changes what you need to publish, and almost nothing about why.
A client called me last spring, worried. Traffic to the site was down, but the phone was ringing about the same as always. Something did not add up.
It added up fine. People were still finding the business. They just were not clicking through to read about it first. They asked an assistant, got an answer, and called the number in that answer.
That is the shift, and it is worth understanding properly before anyone sells you a package to fix it.
What actually changed
For twenty years the deal was straightforward. Someone typed a few words, got ten blue links, and picked one. Your job was to be one of those links, ideally near the top.
Now a large share of questions get answered on the spot. The assistant reads the sources, synthesises them, and hands over a paragraph. The user never sees the ten links. If you were link number four, you did not lose your ranking. You lost the click that used to come with it.
Two things follow from that, and they pull in opposite directions.
The bad news: traffic is a worse measure of visibility than it used to be. You can be more present than ever and see fewer visits.
The good news: being cited in an answer is a stronger endorsement than being link number four ever was. The assistant is not offering options. It is making a recommendation.
An answer has sources
Here is the part that gets lost in the panic. An assistant does not invent a plumber in Nafplio. It reads what exists about plumbers in Nafplio and assembles a reply.
Which means the job did not become mysterious. It became more literal. To be in the answer you have to be a source worth reading: present, consistent, specific and current.
Everything I do for clients on this now comes from that one sentence.
You are not optimising for a ranking algorithm any more. You are trying to be the most useful thing written about your subject.
What actually moves it for a local business
Ordered roughly by how much difference it makes per hour spent.
- Your business profile, filled in completely. Hours, categories, service areas, photos, the lot. It is free, it is structured data, and assistants lean on it heavily. Most businesses fill in half of it and stop.
- The same details everywhere. Name, address and phone identical across your site, your profile, directories and social pages. Three versions of your address is the fastest way to be considered unreliable by something that cross checks sources.
- A real page for each real service. Not one page listing eight services in a bulleted row. A page per service that explains what it involves, what it costs or how pricing works, and who it is for. This is the single most skipped step and the one that gets you quoted.
- Plain answers to the questions you actually get asked. Write down the questions from the phone calls and answer them in normal language on the site. Those are the exact strings people put to an assistant.
- Structure the page so a machine can parse it. Real headings, clean markup, schema for the business, services and reviews. None of this is glamorous and all of it is cheap.
- Reviews, continuously. Not a burst of fifteen in one week two years ago. A steady trickle with actual text in them, because the text is what gets summarised.
- Speed. A slow page is crawled less, read less and abandoned more. Speed has been on this list for a decade and it has never moved down.
Nothing there is a trick. That is the point. The tactics that survived the shift are the ones that were always about being genuinely findable rather than gaming a ranking.
What I stopped doing
Some work that used to pay off now does nothing, and a few things actively hurt.
Writing for keyword density. Repeating a phrase eleven times reads as noise to a system built to understand meaning. Write the sentence a person would say.
Thin location pages. Twenty near identical pages with the town name swapped out. That was always cynical. Now it is also useless, because a summariser has nothing distinct to pull from any of them.
Publishing volume for its own sake. Four thin posts a month lose to one page that genuinely answers something. If you have nothing to say this month, say nothing.
Chasing rank positions. Position three on a query that now resolves in an answer box is worth less than being the source that answer quotes. Different target, different work.
The uncomfortable part
You cannot make a business findable if it has nothing particular to say.
If your site says you offer quality service at competitive prices with a customer focused approach, there is nothing in there to cite. It matches every competitor exactly. No amount of technical work fixes that, because the problem is not technical.
The businesses that do well here are the ones willing to be specific. We serve these areas. We do this and not that. This is what it costs. This is how long it takes. This is what happens if it goes wrong. Specificity is quotable. Vagueness is not.
Measure the thing you actually want
If you take one practical change away from this, make it this one.
Stop reporting on sessions and rank positions. Start counting calls, form submissions, bookings and walk ins that mention finding you online. Those numbers were always the point, and unlike traffic they still mean the same thing they meant three years ago.
My client with the falling traffic ended up fine. We tightened the profile, split the services onto real pages, wrote plain answers to the ten questions the phone kept asking, and stopped looking at the traffic chart. The phone kept ringing, which is what the website was for in the first place.