Does ranking in AI search need separate work, or just good SEO?

Mostly, no - ranking in AI search does not need separate work. The things that get you cited in an AI answer are, to a large degree, the same things that get you ranking well on Google: clear pages, genuine authority, good reviews, and consistent information about who you are and what you do. GEO (Generative Engine Optimisation - optimising to appear in AI-generated answers) is a real thing, but it is being marketed as a brand new discipline with its own price tag far more aggressively than the reality justifies. Let me be precise about where it overlaps with SEO and where it does not.

Why the overlap is so large

AI assistants are not conjuring answers from thin air. They are drawing on the same web the search engines have always crawled, and in Google's case AI Overviews are built right on top of the existing index. So when an assistant decides which business to name, it is leaning on signals that will feel very familiar: is this page clear about what it offers, does this business look trustworthy and established, do other credible sites reference it, is the information about it consistent everywhere it appears.

That is just SEO described in slightly different words. If you have a page that clearly explains a service, that is well-regarded enough to be linked and reviewed, and that says the same thing about your business as your Business Profile and your directory listings do, you have already done most of what GEO asks. There is no secret ingredient sitting behind a paywall.

Getting cited in AI answers mostly comes from the same fundamentals as good SEO. The businesses that were already doing it well tend to show up in AI answers without a special project.

- Tom McSherry

The two levers that matter most

If I had to reduce it to what actually moves whether you get named, it is the same pairing that has always driven ranking: relevance times authority. The page has to clearly be about the thing being asked, and the business has to be credible enough to be worth citing. I have written about why that multiplication matters in relevance times authority, and it applies to AI answers almost unchanged.

• Relevance: is there a clear, well-structured page that directly answers the question a buyer would ask? Vague, everything-page content does badly in AI answers for the same reason it does badly in search.
• Authority: do credible sources reference and review you? An assistant naming three businesses is making a trust judgement, and it leans on the same authority signals search always has.
• Consistency: does your name, service and location read the same across your site, your Business Profile and the directories? Contradictory information makes both Google and an AI assistant less sure of you.

Where GEO genuinely differs

I do not want to pretend there is zero difference, because that is not honest either. There are a few real distinctions worth knowing.

First, measurement is different. You cannot open a rank tracker and see your AI position, because there often is not a position - there is a paragraph that either names you or does not. So the tracking looks different even when the underlying work is the same. Second, the click economics are different: an AI answer can satisfy a reader without ever sending them to you, which changes what content is worth producing. Third, phrasing matters slightly more - content written in plain, direct, question-and-answer language tends to be easier for an assistant to lift and cite. None of that is a new discipline. It is a tilt on the existing one.

So is there a paid 'AI strategy' worth buying?

For most businesses, rarely. If someone is selling you a standalone 'AI SEO strategy' that is separate from, and on top of, your normal SEO, be sceptical and ask exactly what they will do that good SEO does not already cover. Nine times out of ten the honest answer is 'the same fundamentals, plus some tracking'. That does not need to be a separate premium product.

The exception is genuinely research-led, informational businesses - the ones whose buyers do a lot of reading before purchasing - where AI is directly in the buying path and the tracking and content adjustments are worth real, deliberate effort. That is a smaller group than the marketing would suggest, and I have written about exactly who they are in who actually loses traffic to AI search.

My measured take

Do the good SEO. Clear pages, real authority, genuine reviews, consistent information. That gets you most of the way into AI answers without a special project. Layer sensible tracking on top if your buyers are the research-heavy kind. If you are still working out the basics of the landscape, start with what GEO is. And treat anyone selling GEO as a completely separate, expensive discipline with a healthy raised eyebrow. I could be proven wrong if the assistants diverge hard from search over the next couple of years - but on today's evidence, good SEO is most of GEO.

For the local-business version of this answer, see AI search for local business.