How to optimise your website for AI search: a practical checklist

The good news: there’s no secret trick. Most of what makes a site visible to AI is the same groundwork that’s always made it visible to Google. The less good news: plenty of sites are inwittingly getting that groundwork wrong, sometimes without anyone realising.

Here’s a practical checklist we use, in plain English.

How AI search finds its answers

AI search tools don’t just make things up from memory. When someone asks a question, they search the web, pick out a handful of relevant pages, and write an answer based on them, usually with links to the sources.

That means two things have to be true for your business to show up:

1. The AI can find and read your pages. If it can’t get to them, you’re not in the running.
2. Your pages are the best answer. Clear, trustworthy, specific content gets picked over vague or generic content.

Where each tool looks can vary:

Google is refreshingly clear on this. Its own guidance says optimising for AI search is still SEO, and that pages need to be indexed and eligible to show with a snippet to appear in AI Overviews or AI Mode. If you’re invisible in normal search, you’ll almost certainly be invisible in AI search too.

Step 1: Make sure you can be indexed

Everything else depends on this. If search engines haven’t indexed your pages, AI search can’t use them.

• Set up Google Search Console and Bing Webmaster Tools. Most businesses have the first and forget the second. Bing feeds Copilot and is one of the sources behind ChatGPT, so it matters far more than it used to.
• Submit an XML sitemap to both. It’s the simplest way to tell them which pages exist.
• Check the indexing reports. Look for important pages marked as “crawled but not indexed”, “excluded by noindex” or blocked by robots.txt. A stray noindex tag left over from a site launch is more common than you’d think.
• Don’t hide your content from snippets. Google needs to be able to show a snippet of a page for it to appear in AI features, so avoid nosnippet tags on pages you want found.
• Check your AI settings in Search Console. Google now lets site owners control whether their site appears in its generative AI features, so make sure yours hasn’t been switched off.

Both tools now include AI reporting too: Search Console has a generative AI performance report, and Bing Webmaster Tools has an AI Performance report. They’re the closest thing to a free scoreboard for AI visibility.

Step 2: Let the right AI crawlers in

AI companies use their own crawlers, and your robots.txt file tells them what they’re allowed to read. The catch is that each company has separate crawlers for separate jobs:

This gives you a real choice. Many businesses are happy to be cited in AI answers but would rather their content wasn’t used to train AI models. You can block the training crawlers and still allow the search ones. Block them all, and you’re likely to disappear from those AI answers.

Watch out for your security settings. Even with a perfect robots.txt, a firewall or CDN can block AI crawlers before they reach it. Cloudflare, which sits in front of a large share of websites, has blocked AI crawlers by default on new domains since July 2025 and has refined those defaults since. If your site uses Cloudflare or a security plugin, check what it’s set to. It’s entirely possible to be paying for SEO while your firewall is turning the AI crawlers away.

Step 3: Make sure machines can read your pages

When you visit a website, your browser runs a lot of code to build the page you see. Search engines and AI crawlers don’t always do the same.

Google can process JavaScript, but it’s more work and can be slower. Many AI crawlers are less patient: if your main content only appears after JavaScript runs, there’s a real risk they see a mostly empty page.

This matters most for:

• Sites built as JavaScript apps (React, Vue and similar), including most sites made with AI website builders
• Content loaded in tabs, accordions or “load more” buttons that only appear after a click
• Prices, reviews or product details pulled in by a third-party widget

A quick test: view the page source in your browser (not the inspector, the raw source). If your headings and main text aren’t in there, a crawler may not see them either. The fix is usually server-side rendering or pre-rendering, so the important content is in the HTML from the start.

While you’re there, check the basics:

• one clear H1
• logical headings
• descriptive page titles
• …and a site that loads quickly and works well on mobile.

Step 4: Add schema markup

Schema markup (also called structured data) is code that labels the information on a page in a way machines understand: this is a business, this is its address, this is a product, this is its price, this person wrote this article.

Let’s be clear about what it does and doesn’t do. Google says schema isn’t required for its AI features and there’s no special “AI schema” to add. But it still recommends using it, because it makes you eligible for rich results in normal search, and it removes guesswork about who you are and what you offer. That clarity helps every system that reads your site, AI included.

The types most businesses should have:

Three rules keep it working:

• Only mark up what’s visible on the page. Schema that describes content visitors can’t see breaks Google’s guidelines.
• Keep it consistent. Your business name, address and phone number should match everywhere: schema, footer, Google Business Profile and directories.
• Test it. Run key pages through Google’s Rich Results Test and the Schema.org validator after any change.

Step 5: Write content that AI can quote

Once an AI can read your site, the question becomes whether it wants to use it. This is where most of the long-term gains are.

• Answer the question early. Put the direct answer near the top of the page or section, then explain. If someone asks “how much does a conservatory roof replacement cost?”, say so plainly before the detail.
• Use clear headings that match real questions. Headings help readers scan, and they help AI find the relevant part of a page.
• Be specific. Prices, timescales, locations, materials, processes. Vague marketing copy gives an AI nothing to cite.
• Show your expertise. Put real names on articles, explain who you are and why you know this, and include first-hand experience: case studies, real examples, your own photos.
• Say something others don’t. Google’s guidance warns against “commodity” content that just repeats common knowledge. Your own process, opinions and results are what make a page worth citing.
• Keep it current. Update key pages when prices, services or facts change, and show when they were last updated.

What you don’t need to do is rewrite everything “for AI”. Google specifically says there’s no need to chop content into tiny chunks or write in a special style. Write well for people, and you’re writing well for AI.

Step 6: Build trust beyond your own site

AI tools don’t only look at your website. They also look at what the rest of the web says about you, especially for questions like “who’s the best electrician in Northampton?”

• Google Business Profile. For local businesses, this is essential. Keep categories, services, hours and photos up to date, and respond to reviews. Google says its AI answers can draw on business profile information directly.
• Merchant Center. For online shops, a product feed in Google Merchant Center helps your products appear in AI shopping answers as well as normal search.
• Reviews. Genuine reviews on Google and relevant review sites shape how AI describes you.
• Consistent listings. Make sure directories, trade bodies and social profiles show the same name, address, phone number and description.
• Real mentions. Local press, industry publications, partner websites, podcasts and community forums all count. Google warns that chasing fake or paid-for mentions doesn’t work, so focus on earning genuine ones.

What about llms.txt?

You may have been told you need an llms.txt file: a plain text file at the root of your site that gives AI tools a tidy summary of your most important pages. It’s become a popular “AI SEO” recommendation.

The honest answer: it’s optional, and it won’t move the dial on its own.

• Google’s own AI search guidance lists llms.txt among the things you can ignore, and says Search doesn’t use it at all.
• No major AI company has publicly committed to using it to decide what to cite.
• Large studies of server logs in 2026 found the vast majority of llms.txt files receive little or no traffic from AI crawlers.

It does no harm, and it can be useful for software and documentation sites where AI coding tools do read it. What is interesting, is that on Google’s PageSpeed metrics (https://pagespeed.web.dev/) llms.tt is cited as being ready for agentic search.

Your AI search readiness checklist

• Google Search Console and Bing Webmaster Tools set up and verified
• XML sitemap submitted to both
• Key pages indexed, with no stray noindex or nosnippet tags
• Site included in Google’s generative AI features in Search Console
• robots.txt allows AI search crawlers (and blocks training crawlers, if you prefer)
• Firewall, CDN and security plugins not blocking AI crawlers
• Main content visible in the page source, not only after JavaScript runs
• Organization or LocalBusiness schema in place, plus Product, Article and Breadcrumb schema where relevant
• Schema tested and matching what’s visible on the page
• Key pages answer the main question near the top, with clear headings
• Articles have named authors with real expertise
• Google Business Profile complete and up to date
• Business details consistent across the web
• AI performance reports checked monthly in Search Console and Bing Webmaster Tools

The short version

AI search hasn’t replaced the basics. It’s made them matter more. If your site can be found, read and trusted, and your content genuinely answers the questions your customers ask, you’re in a strong position for ChatGPT, Google’s AI features and whatever comes next.

Most of the problems we find aren’t complicated. They’re a blocked crawler, a missing Bing setup, schema that was never added, or content that’s hidden behind JavaScript. Small fixes, but they can be the difference between being cited and being invisible.

Not sure where your site stands? Get in touch and we’ll run through the checklist with you, tell you what’s working, and fix what isn’t.

Sources

• Optimizing your website for generative AI features on Google Search (Google Search Central, updated July 2026)
• AI crawlers in robots.txt (Scalableloops, September 2026)
• ChatGPT built its own search index (Peec AI, September 2026)
• We publish an llms.txt. The 2026 data says almost nothing reads it (AI SaaS Radar, September 2026)