How to Audit Your Website for AI Search Readiness

Your website probably isn't ready for AI search. Here's how to find out.

Right, so I've been running these audits for businesses around Hitchin and Stevenage for the last 18 months, and there's a pattern. Most websites were built for Google in 2019. They're still optimised for how search worked back then. Which means they're basically invisible to ChatGPT, Perplexity, and whatever else people are using to find local services these days.

And the business owners don't know. Because their Google Analytics still shows traffic. Because they're still getting some leads. But when I show them what ChatGPT says when someone asks "who should I use for X in Baldock", their business just... isn't mentioned. At all.

So this is how you audit your own site. Properly. Not the surface stuff. The actual things that determine whether AI search engines can understand, trust, and recommend your business.

Start with the question test

Open ChatGPT. Or Claude. Or Perplexity if you're fancy. Then ask it the exact questions your customers would ask.

"Who should I hire for [your service] in [your town]?"

"What's the best [your service] company near Letchworth?"

"I need [specific problem you solve] in North Hertfordshire, who do you recommend?"

Write down what it says. Does it mention you? Does it mention your competitors? Does it just give generic advice and then tell you to Google it?

I did this with a client in Royston last month. Roofing company. Been in business 30 years. ChatGPT had never heard of them. But it recommended three of their competitors. Why? Because those competitors had their information structured in ways that AI can actually parse and understand.

This isn't about gaming the system. It's about making your information accessible to how people are actually searching now.

Check if your site speaks structured data

AI search engines don't read your website the way humans do. They look for structured data. Schema markup. The machine-readable layer underneath your pretty design.

Go to your homepage. Right-click, view source. Then Ctrl+F and search for "schema" or "application/ld+json".

If you find nothing, you've got work to do.

If you find something, copy it and paste it into Google's Schema Markup Validator. Does it actually validate? Is it the right type of schema for your business? LocalBusiness, Service, FAQPage, whatever's relevant.

Most websites I audit have either no schema at all, or they've got some half-arsed implementation from a plugin that doesn't actually tell AI engines anything useful. It'll say you're a business at an address, but it won't explain what you actually do, who you serve, what problems you solve.

The depth matters now. You need schema that says "we're a plumber serving these postcodes, handling these specific issues, with these qualifications, and here's proof of our reviews". That's the level AI engines are looking for.

Your content needs to answer complete questions

Look at your service pages. Are they written like brochures or like answers?

Because here's what happens. Someone asks an AI "how much does [your service] cost in Stevenage?" The AI goes looking for comprehensive answers. If your pricing page just says "contact us for a quote", you're out of the running. Even if your pricing is competitive.

The AI will find your competitor who wrote "typical projects range from £X to £Y depending on these factors" and recommend them instead. Not because they're better. Because they actually answered the bloody question.

Go through your main pages. For each one, write down the questions customers actually ask about that topic. Then see if your page genuinely answers those questions. Completely. With specifics.

Not marketing fluff. Not "we pride ourselves on quality". Actual information that helps someone make a decision.

I see this most with FAQs. Businesses have these token FAQ sections with five generic questions. "Do you offer free quotes?" Well yeah, everyone does. That's not what people are asking AI. They're asking complex, specific stuff. "What should I do if I find damp in my loft in winter?" That's the level you need to be writing at.

Test your local signal strength

AI search is heavily weighted toward local intent. But your website needs to make your local coverage crystal clear.

Don't just mention "Hertfordshire" once in your footer. You need proper location pages for each area you serve. Or at minimum, clear signals throughout your content about where you actually work.

Check this: search your own website (site:yourwebsite.co.uk) and add one of your target towns. How many results? What pages mention that location? Is it just in a list somewhere, or is there actual content about serving that area?

Because when someone in Letchworth asks AI for recommendations, the engines look for businesses that demonstrably serve Letchworth. Not businesses that theoretically might serve anywhere in Hertfordshire if you call them.

You need pages that talk about the specific needs of customers in those areas. Local landmarks, local issues, evidence you actually work there. Case studies from that town. That sort of thing.

Look at your evidence layer

AI engines are paranoid about recommendations. They want proof you're legitimate. That means they're looking for:

• Reviews on third-party platforms (Google, Trustpilot, whatever's relevant for your industry)
• Professional accreditations and memberships
• Case studies with actual details and outcomes
• Clear information about who runs the business and their qualifications

Go through your site. Is this information there? Is it detailed? Or is it just logos of accreditation bodies with no context?

I audited a site last week that had "Accredited by [Industry Body]" in the footer. But no link. No explanation of what that accreditation means. No date. From an AI's perspective, that's worthless. It can't verify it, so it ignores it.

You need verifiable, specific evidence. "We're Gas Safe registered, number 123456, you can verify this here". "We've been members of [Association] since 2015, here's our profile". That level of detail.

Check if you're citation-worthy

Here's one that catches people out. AI engines don't just look at your website. They look at what other sources say about you.

Google your business name. What comes up? Your website, sure. But what else? Local directories? Review sites? News mentions? Industry listings?

Because AI search uses citation signals. If you're mentioned on trusted sites, in directories, in local business listings, that builds authority. If you only exist on your own website, you're basically invisible to AI.

Make a list of where you're currently listed. Then check the big ones you're missing. Google Business Profile is obvious. But also industry-specific directories, local business associations, Yell if that's still relevant for your industry (it is for some, isn't for others).

The consistency matters too. Is your NAP (name, address, phone) the same everywhere? Or did you move office two years ago and half your listings still show the old address? AI engines see that inconsistency and trust you less.

The mobile structure test

Pull up your site on your phone. Not to see if it looks pretty. To see if the information hierarchy makes sense.

Because AI engines are increasingly parsing websites the way a screen reader would. They look at heading structure, content order, how information is organised. If your mobile site hides half your content behind hamburger menus and accordion tabs, that content is weighted less.

Check your headings. View source again, look for H1, H2, H3 tags. Do they tell a logical story? Or is your H1 just your logo, and everything else is in divs with CSS that makes them look like headings?

This sounds technical but it matters. AI engines use heading structure to understand what your page is about and how information relates to each other. If your headings are a mess, they can't parse your content properly.

What happens when you actually audit this stuff

Most businesses find they're failing on at least four of these areas. Usually more.

And look, that's not a criticism of whoever built your website. They built it for 2019. It probably worked great for 2019. But it's 2026 now and people are finding businesses differently.

The good news is most of this is fixable. It's not like you need to rebuild everything from scratch. You need to add structure, answer questions properly, strengthen your local and authority signals.

If you're in North Hertfordshire and you want someone to actually audit this properly, not just run your URL through some automated tool, we do that. Book a call and I'll tell you straight what's missing and what matters most for your specific business. Or if you want to see what proper AEO in North Hertfordshire looks like, that page has more context.

But honestly, even if you just work through this list yourself, you'll be ahead of 80% of local businesses who still think SEO means cramming keywords into page titles.