When someone asks an AI system a question about a business like yours, does your business show up in the answer? That's the real diagnostic question behind AEO and GEO — Answer Engine Optimization and Generative Engine Optimization — and it's the single most underdiagnosed blind spot I find in SME businesses right now. Here is the five-step check I run, in order, on every audit.
Most of the SME owners I sit down with have already "done something" with AI — a chatbot, a content tool, an AI line added to the website copy. That's usually not where the real gap is. Search behaviour has quietly shifted: a growing share of people who'd once have Googled "plumber near me" or "dental practice Milton Keynes" are now asking ChatGPT, Perplexity, or a Google AI Overview instead — and getting a confident, specific answer without ever seeing a list of blue links to click through. If your business isn't part of that answer, you don't lose a click. You lose the moment the decision gets made.
The check I run, in order
The direct-question test
I open three or four AI tools — ChatGPT, Perplexity, Gemini, and a Google AI Overview where one appears — and ask the exact questions a real customer would ask. Not branded searches ("what is [company name]"), but the questions people actually type when they don't yet know who to hire: "best AI consultant for SMEs in Milton Keynes," "who audits AI readiness for small businesses in the UK." If the client doesn't appear in any of these answers, that's the headline finding, and usually the moment the conversation gets serious.
Structured data and schema
AI systems don't read a website the way a human does — they read the structured data underneath it. Most SME sites have none, or have generic boilerplate schema left over from a template. I check for Organization schema (so AI tools can confidently resolve who the business actually is), FAQPage schema on any page answering real customer questions, and, where relevant, Service or Product schema. This is usually the single fastest fix on the list — it's technical, not creative, and it can often be added without touching the site's design at all.
Entity consistency across the web
AI engines build confidence in an answer by cross-referencing multiple sources. If your business name, address, credentials, and description say slightly different things on your website, your LinkedIn page, and your Google Business Profile, that inconsistency quietly erodes how confidently an AI system will cite you. I check name/description/credential consistency across every property the business controls, and flag the gaps.
Content that actually answers questions
Most SME websites are written to describe the business, not to answer the questions a prospective customer — or an AI system summarising an answer for them — actually has. I look for whether the site has content that directly and plainly answers the questions I asked in step one, in the customer's own language, not marketing language. If it doesn't, that's usually the biggest content gap, not a technical one.
Crawler access
A smaller technical check, but a real one: whether the site's robots.txt and llms.txt actually allow AI crawlers — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, CCBot — to access the site at all. I've seen sites that inadvertently block every one of them. If a crawler can't read the site, nothing else on this list matters.
What this actually catches
On a recent audit, a client's business was structurally sound — decent website, reasonable SEO — but appeared in zero of the direct-question tests across all three AI tools. The cause, once we looked, was almost entirely steps two and three: no schema markup at all, and three different versions of the business address across their own site, Google Business Profile, and a partner listing. Neither fix required a rebuild. Both were done inside two weeks.
It matters now. Roughly a quarter of Google searches already surface an AI Overview instead of, or before, a list of links. That share is only going up.