How AI SEO works
How AI answers pick their sources
When someone asks an AI assistant “who’s the best orthodontist in Leeds?” or “what should I look for in a roofing contractor?”, the tool typically does two things. It searches the web for relevant pages, and it combines what it finds with what it already learned in training. The businesses it names are the ones that show up clearly and consistently in both places.
That has a few practical consequences.
- You need to be findable in search. Tools like Google AI Overviews and ChatGPT search lean heavily on pages that rank well. If you don’t rank, you’re far less likely to be cited. That’s why AI SEO usually starts with classic SEO foundations.
- You need to be easy to quote. A page that answers the question clearly, near the top, in a sentence or two, is easier to lift into an answer than one that buries the point.
- You need to be talked about elsewhere. A business mentioned on industry sites, local press, directories and review platforms looks more established than one that only describes itself.
Why brand mentions on other sites matter more than ever
In classic SEO, links are a strong ranking signal. In AI search, mentions count too, even unlinked ones. When an AI tool sees your business named across trusted sources, all saying the same things about what you do and where, it becomes more confident including you.
This is what entity optimization means in practice. Your business is an “entity” with facts attached: name, services, location, founder, years in business, the areas you serve. If those facts differ between your website, your Google Business Profile, your LinkedIn page and three directories, you look less certain. Cleaning that up is unglamorous work, but it’s one of the most reliable things you can do. It also overlaps with the citation work in local SEO, so local businesses often get both from one effort.
What answer-first content looks like
Most service pages are written to persuade. They open with a slogan, build up slowly and get to the facts eventually. AI tools, and busy readers, want the opposite.
Answer-first content puts the direct answer at the top of each section, then adds the detail. For example:
- Weak: “At Smith Dental we believe every smile deserves the very best care, which is why…”
- Strong: “Invisalign at Smith Dental costs between X and Y and usually takes 12 to 18 months. Here’s what affects the price.”
The strong version can be quoted on its own. It answers the question a buyer actually asked. You still persuade, but you do it after you’ve been useful.
FAQs are a natural fit for this style. Each question should be one a real customer asks, answered in 40 to 80 words that make sense without the rest of the page.
Structured data that helps
Structured data (schema markup) is code that states facts about your page in a format machines read reliably. The types that matter most for AI visibility are usually:
- Organization or LocalBusiness, with your name, logo, contact details, address and links to your profiles elsewhere.
- Person, for the founder or key experts, connected to the business.
- Service, describing each service and who provides it.
- FAQPage, matching the questions visibly answered on the page.
- BreadcrumbList, showing how pages relate to each other.
Schema won’t make a weak page rank, but it removes ambiguity about who you are and what you do. Getting it right across a whole site is part of technical SEO, and it’s one of the first things I fix.
What’s hype and what’s measurable
AI search attracts a lot of overpromising. Here’s an honest split.
Measurable:
- Whether you’re mentioned for a fixed set of prompts, tracked monthly across the main AI tools.
- Referral traffic from AI tools in your analytics.
- Your organic rankings and impressions in Search Console, which still drive much of what AI tools see.
- How consistent your brand facts are across the web.
Hype:
- “Guaranteed ChatGPT rankings.” AI answers vary between conversations, and nobody controls them.
- Secret prompts or tricks that “force” a mention. They don’t last, if they work at all.
- Treating llms.txt as a ranking factor. It’s cheap to add and worth doing, but there’s no evidence that major AI tools rely on it.
The useful way to think about AI SEO is that it rewards the same things good marketing always has: being clear, being consistent and being mentioned by people other than yourself. The difference is that AI tools surface that quickly, and in a format where only a few names get shown.
If you want to know whether AI tools mention your business today, tell me about it below. I’ll come back with a fixed quote for testing your visibility and improving it.