There is a quiet revolution happening in search, and most small businesses have no idea it is taking place. When someone asks ChatGPT "who builds websites on the Wirral?" or "best locksmith near me", the answer does not come from a Google results page. It comes from a language model that has already decided which businesses are worth mentioning and which are not.
If your website was not part of that decision, you are invisible. Not buried on page five. Not competing for a click. Simply absent from the conversation entirely.
How AI Search Actually Works
Large language models like ChatGPT, Claude, and Perplexity do not crawl the web in real time the way Google does. They are trained on enormous datasets, and increasingly they use retrieval-augmented generation (RAG) to pull in live web content when answering questions. But they are selective. They favour content that is structured, citable, and semantically clear.
That means your beautifully designed homepage with a hero image and a tagline like "We Build Digital Experiences" tells an AI absolutely nothing useful. There is no structured data to extract. No clear statements of what you do, where you do it, or who you do it for.
The Four Things AI Search Engines Look For
1. Structured Data (Schema Markup)
Schema markup is machine-readable metadata embedded in your HTML. It tells search engines and AI models exactly what your business is, where it operates, what services it offers, and what customers say about it. Without it, an AI model has to guess what your page is about from raw text, and it usually guesses wrong or ignores you entirely.
2. An llms.txt File
This is the robots.txt equivalent for AI. It tells language models what your site is about, what content they are allowed to reference, and how to cite you. Most businesses have never heard of it. The ones that have it are already gaining an edge in AI-generated recommendations.
3. FAQ Sections With Real Answers
AI models love question-and-answer pairs because they map directly to how users query them. A well-written FAQ section does not just help human visitors. It gives AI models citable, structured answers they can surface in response to natural language questions.
4. Citable, Specific Content
Vague marketing copy is useless to an AI. Statements like "We have over 10 years of experience delivering bespoke solutions" give a model nothing to work with. Compare that with "Bright Loop Media builds custom websites for small businesses across the Wirral, Liverpool, and Chester, typically delivering a live site within 48 hours." The second version is specific, factual, and citable.
What You Can Do Today
Start by auditing your site with a GEO lens. Check whether you have LocalBusiness schema markup. Add an llms.txt file to your root domain. Write an FAQ section that answers the exact questions your customers ask before they hire you. Rewrite your service descriptions to be specific and factual rather than aspirational.
If you are not sure where to start, we offer a full GEO audit that scores your site across all the factors AI search engines care about and gives you a prioritised action plan. The businesses that move on this now will be the ones AI recommends in six months. The rest will still be wondering why their traffic is declining despite "good SEO".
Traditional SEO is not dead, but it is no longer enough. The question is not whether AI search will replace Google for your customers. It is whether you will be visible when it does.
Read next: our pillar guide on Generative Engine Optimisation covers the full 12-point GEO audit, the schema patterns that matter, and a worked llms.txt example. The complete version of what this article touches on.
