How AI Search Works and Why Your Business Is Invisible to It

When a potential customer asks ChatGPT to recommend a WordPress developer in the UK, your name should come up. For most businesses, it doesn’t — and the reason isn’t that AI doesn’t know you exist. It’s that AI can’t trust you enough to recommend you.

Here’s how that trust is built, and what’s missing from most sites.

How AI assistants actually answer questions

ChatGPT, Perplexity, and Google’s AI Overviews don’t work like traditional search engines. They don’t rank pages — they synthesise answers from sources they’ve indexed, weighted by how authoritative and consistent those sources appear.

When they’re recommending a business, they look for:

• Corroboration — the same entity (name, URL, location) mentioned consistently across multiple sources
• Structured data — machine-readable markup that confirms what your business does, where it operates, and who it serves
• Crawlable content — pages that answer the specific questions users are likely to ask

If your site doesn’t have all three, you’ll be skipped in favour of someone who does.

The structured data gap

Most WordPress sites either have no structured data or generic schema added by an SEO plugin without customisation. A generic Organization schema that just names your business does almost nothing.

What actually helps is specific, accurate schema for your services, your location, your reviews, and your person or business entity. The more signals AI can verify, the more confident it becomes in citing you.

Crawlability still matters — more than ever

Google bots and AI crawlers behave differently. AI crawlers like GPTBot and Claude’s crawler look for specific signals: llms.txt files, clean sitemaps, logical content architecture, and pages that directly answer common questions in your field.

If your robots.txt is blocking them (which many WordPress security plugins do by default), you’re invisible to AI before the conversation even starts.

The content structure problem

AI assistants are pattern-matching engines. They’re very good at extracting answers from well-structured content — clear headings, direct answers to specific questions, factual claims that can be corroborated elsewhere.

Most business websites are written for humans skimming a sales page. That’s useful, but it’s not the format AI excels at parsing. The fix isn’t rewriting your entire site — it’s adding a layer of structured, question-answering content that runs alongside your existing copy.

Where to start

Run your site through Google’s Rich Results Test to see what structured data you have. Then search for your business name in ChatGPT and Perplexity and see what comes back. If the answer is nothing, or wrong information, you have a corroboration problem that structured data alone won’t fix.

That’s what a GEO audit is for.