Try something. Open whichever AI assistant you have on your phone and ask it to recommend a business that does what you do, in your town. Then ask it what it knows about your business by name.
I do this every Monday for four websites, including my own, as part of a search audit that runs before I have finished my first coffee. When I started running these checks, the assistants were getting basic facts wrong about more than one business: an old service still listed as current, a competitor recommended first, a town that was not quite right. Nobody had lied to them. They had read an old page, found nothing newer, and filled in the gaps.
That is the new version of the Google problem, and most small businesses have not noticed it yet. A growing share of people, particularly younger ones and particularly for services, ask an assistant rather than search. And the assistant does not give ten blue links. It gives one answer, and if you are not in it, you are not anywhere.
The good news is that what decides the answer is not clever. It is five boring things, and they are all within your control.
The first is whether your website states plainly what you do, where, and for whom. Not in a headline written to sound impressive. In a sentence a machine can quote. “NE Support is run by Louise Edwards, an AI and business support specialist based in Durham.” That sentence is going on my site for exactly this reason.
The second is whether the facts agree everywhere. Your website, your Google listing, your LinkedIn, your Facebook page. If one says Durham and one says Newcastle, if one says you do kitchens and one says you stopped, the assistant hedges or picks one. Pick one yourself.
The third is whether you answer questions. Assistants love a page that asks “how much does this cost” and then answers it. Most business websites are allergic to the question. A short set of plain questions and answers on your site, about price, area, timescales and what you will not do, is worth more than any amount of clever copy.
The fourth is reviews, and specifically recent ones. An assistant asked to recommend someone leans heavily on Google reviews, and it notices dates. Forty reviews from 2022 read as a business that has gone quiet. Four from last month read as alive.
The fifth is whether anyone else mentions you. A local directory, a supplier’s website, a chamber of commerce listing, a mention in the local paper. Assistants trust what other sites say about you more than what you say about yourself, same as people do.
None of this is a project. It is an afternoon, and then a check every so often to make sure the answer has not drifted. It is also the kind of thing I build into the monthly report for clients, because if a machine is going to describe your business to a stranger, you should at least know what it is saying.
Go and ask it. If you do not like the answer, that is fixable, and it is fixable faster than you think. If you would rather I did the checking, the Business Systems and AI Audit starts there, and the 3 Quick Wins post is a good place to start on your own.