What is AI search, in plain English?
Asking a question and getting a written answer that names two or three businesses, with the links underneath. Google now does it at the top of many results pages, which it calls an AI Overview, and ChatGPT, Gemini and Perplexity do it as a conversation. Somebody in Halifax can type who is a good commercial electrician near me and be handed names, and if yours is not one of them the search is over before you were seen.
The work of getting named is called generative engine optimisation, or GEO, spelt generative engine optimization in the American sources, and sometimes AI SEO or answer engine optimisation. The names matter less than the method.
Is AI search different from SEO?
It builds on SEO, and a business that is invisible on Google will be invisible to the assistants too. They draw on the same search indexes, the same Google Business Profile and the same reviews, and then on everything written about you elsewhere, so the foundations are identical. What changes is how much the assistants care about clarity. They need to work out, from sentences, that you are one business, in one place, doing named things.
Dolphin Heating in Huddersfield is a one-van gas engineer whose site says plainly what Fay does and where. She is number one on Google for female gas engineer Huddersfield and named in the AI answer for it, on clarity, with no size behind her at all.
What information does an assistant need about my business?
Who you are, what you do, where you do it, who you have done it for and what other people say about it, all written in ordinary sentences somewhere it can read. In practice that means a page per service, the towns you cover named in prose, an about page with real people on it, case studies with the client and the place named, and the same business name, address and phone number everywhere from the website to the Google profile to the trade directories.
- A plain first line: the business, the town and the service in one sentence
- A page for each service, saying who it is for and where
- Named projects and clients, with photographs you took yourself
- Google reviews with real names on them
- Structured data behind the page that repeats the same facts
- The same name, address and phone number on every listing
Do reviews and case studies matter for AI visibility?
Yes, they are the evidence an assistant uses to decide you are credible enough to name. A model asked for a recommendation is looking for corroboration: reviews on Google, projects written up with a client and a place attached, video of the work, and mentions on sites that are not yours. Bond Brickwork in Brighouse are recommended by AI search for commercial brickwork, and the record behind that is years of named projects, drone footage and search work, all of it public.
Third-party mentions are the part most local businesses have least of. Trade bodies, suppliers, local press and the organisations you sponsor all count, which is one reason digital PR has become part of the marketing.
How do I check what AI says about my business now?
Ask it the question a customer would ask, several ways, and write down who it names. Try Google with a question that triggers an AI Overview, then ChatGPT, Gemini and Perplexity with the same question: who is a good commercial flooring contractor in West Yorkshire, which electrician in Halifax installs EV chargers, and so on. Note who is named, who is missing and what the answer says about each. Repeat it monthly. That list of questions is how we measure AI visibility for clients, because a search position tells you nothing about whether the assistant mentioned you.
What would Halizona do for a Halifax business?
Sort the search foundations first, then make every fact about the business explicit, then build the proof and the mentions, and measure it against the questions your customers ask. Nothing on that list is exotic. It is the same find, trust and choose work the rest of this site describes, done carefully enough that a machine can read it back.
