A January 2026 analysis of one million websites has found that only 0.3% have implemented AI discovery files. That leaves 997,000 sites invisible to the AI systems now answering 2.5 billion queries daily.
Not a typo. Less than one percent.
We ran our AI Visibility Checker across 847 UK business websites this week. The average score was 12 out of 100. Most had zero AI discovery files in place.
ChatGPT processes those 2.5 billion prompts from 800 million weekly active users. Google's AI-powered search now generates direct answers, without any website click, in 58% of searches. If your website lacks AI discovery files, you're not competing for those queries. You simply don't exist.
What the Research Actually Shows
The data comes from a scan of one million websites conducted in early January 2026. Researchers looked for the presence of AI discovery files: the structured text documents that help AI systems understand what a business does.
Only 3,000 websites out of a million had them. 0.3%.
The timing matters. According to McKinsey's latest AI adoption research, 78% of enterprises now use AI in at least one business function. These systems need information sources they can process quickly. Websites without AI discovery files aren't those sources.
"Large language models increasingly rely on website information, but face a critical limitation: context windows are too small to handle most websites in their entirety."
llms.txt specification, llmstxt.org
That quote from the spec itself spells out the core problem. When I first read it, I thought: this is exactly what we see with our hosting customers. They build beautiful sites, invest in SEO, then wonder why ChatGPT can't describe their business correctly. The AI isn't broken. It just can't read 47 pages of WordPress content in one go. It needs a summary. That's what llms.txt provides.
Why AI Systems Can't Find Most Websites
Here's what happens when ChatGPT or Claude tries to understand your website without AI discovery files.
The AI hits a hard technical limit. Context windows can only hold so much text. Even the most advanced models can't parse an entire website with its styling, scripts, navigation menus, sidebar widgets, footer content, and actual page copy all mixed together.
Without structured guidance, the AI samples fragments, makes assumptions, and often gets things wrong. We see this constantly with our AI Visibility Checker. Businesses assume ChatGPT reads their entire website like a human would. It doesn't.
AI discovery files solve this by providing a structured map. They tell the AI: here's who we are, what we do, what we're called, and how to describe us accurately. We've written a full breakdown of how each file works and why this pattern mirrors robots.txt and sitemaps.
The 10 Files AI Systems Actually Look For
A complete AI Site Identity includes ten files, each serving a specific purpose:
| File | Purpose |
|---|---|
| llms.txt | Core business identity in markdown. Your AI-readable elevator pitch telling AI systems exactly what your business does. |
| llm.txt | Compatibility variant that redirects to llms.txt. Catches AI systems requesting the singular filename. |
| llms.html | Human-readable visual representation. A webpage version of your AI identity with Schema.org structured data. |
| ai.txt | Operational parameters and permissions. Specifies how AI systems can use your information and what topics to avoid. |
| ai.json | Machine-readable interaction guidance with JSON Schema validation. The structured equivalent of ai.txt for systems that parse JSON natively. |
| identity.json | Structured canonical data in JSON. Machine-readable business information with zero ambiguity. |
| brand.txt | Identity consistency rules. Prevents AI from inventing incorrect names or descriptions for your business. |
| faq-ai.txt | Common queries optimised for AI. Provides verified answers when someone asks ChatGPT about your business. |
| developer-ai.txt | Technical documentation for software queries. Accurate specs when someone asks about your tech stack. |
| robots-ai.txt | AI-specific crawler permissions. Works alongside traditional robots.txt for AI crawling behaviours. |
Our complete guide to all 10 AI discovery files explains each one in detail. Most businesses start with llms.txt and ai.txt, then build from there.
Why This Creates a Competitive Advantage
Think about those numbers again. 0.3% of websites have AI discovery files. That means 99.7% don't.
When someone asks ChatGPT "recommend a good WordPress hosting company in the UK," the AI looks for businesses it can confidently describe. Websites without AI discovery files get skipped. Not because they're bad, but because the AI can't efficiently understand them.
"If your content isn't in the Docs section of an llms.txt file, your brand and your narrative will not be found."
Jeremy Nguyen, Swinburne University academic and AI researcher
Nguyen's warning is blunt, and from what we've seen across hundreds of sites, he's right. After implementing AI discovery files on our own WordPress hosting platform, we tracked a clear improvement in how AI systems cited us. When ChatGPT describes UK WordPress hosting options now, 365i appears accurately. Because we told it exactly what we are.
The gap widens daily. Early adopters sit in that 0.3% while competitors remain invisible. By the time AI discovery becomes standard practice, the advantage window closes.
The WordPress Problem Nobody Talks About
WordPress powers 43% of all websites. Most WordPress users focus on traditional SEO: keywords, backlinks, page speed, Core Web Vitals. None of that helps AI systems understand your business.
The problem is worse for WordPress sites specifically. Plugin-heavy codebases, complex page structures, dynamic content generation. All of it creates barriers for AI parsing. When ChatGPT tries to understand a typical WordPress site, it encounters navigation menus, sidebar widgets, footer links, and scattered content blocks.
Without AI discovery files providing clear structure, the AI guesses. Often incorrectly. A free WordPress plugin now handles all 10 files automatically, which removes the biggest barrier for WordPress site owners.
We manage thousands of WordPress sites on our WordPress Turbo Hosting platform. The pattern is consistent: sites with AI discovery files get cited accurately by AI systems. Sites without them get ignored or misrepresented.
This isn't about SEO rankings. It's about whether AI systems can recommend you at all. We covered this distinction in detail in our post about why AI discovery files aren't SEO.
Generate AI Discovery Files from your dashboard
Using WordPress? Install the plugin and create all 10 files in minutes. No coding, no configuration files to edit manually.
Get the Plugin →DIY vs Professional Setup
Two options. Here's what each involves:
| Aspect | DIY | Professional (£295 + VAT) |
|---|---|---|
| Time | 4-8 hours minimum | 24-48 hours delivery |
| Technical knowledge | Markdown, JSON, file placement | None required |
| Content audit | You find contradictions | Included |
| Validation | Use free checker | Full testing included |
| Risk | Formatting errors break everything | Issues fixed before delivery |
For the DIY route, our llms.txt and ai.txt creation guide walks through the technical process step by step.
Where Do You Stand?
The 0.3% figure won't last. As awareness grows, AI discovery files will become standard, just like SSL certificates and mobile responsiveness before them.
Right now, that 99.7% gap is pure opportunity. But only if you act while competitors remain invisible.
Our free AI Visibility Checker analyses your website in 30 seconds. No signup required. You'll get a score out of 100 plus specific recommendations.
The average UK business scores 12. Sites with proper AI discovery files score 80+.
Check where you stand. Then decide whether to close the gap yourself or understand why ChatGPT can't find you first.
Frequently Asked Questions
What exactly are AI discovery files?
Structured text documents placed on your website that help AI systems like ChatGPT, Claude, and Google AI understand your business accurately. They include files like llms.txt for core identity, ai.txt for usage permissions, and brand.txt for naming consistency. Without them, AI systems guess based on fragments of your site content.
Why do only 0.3% of websites have AI discovery files?
Most website owners don't know they exist. The llms.txt specification only gained traction in late 2025, and traditional SEO practices don't cover AI visibility. This creates a window of competitive advantage for early adopters.
How is this different from SEO?
SEO optimises for search engine rankings to drive website traffic. AI discovery files ensure AI systems understand and describe your business accurately when answering questions directly, often without any website click. You need both.
Will AI discovery files guarantee ChatGPT mentions my business?
No. AI systems make their own decisions about relevance and quality. What the files guarantee is that when AI does reference your business, it works from accurate information rather than guessing. You control the narrative, not the mentions.
How long does it take to implement AI discovery files?
DIY takes 4-8 hours if you understand the file specifications and have your business information organised. Professional setup through our AI Site Identity service delivers all 10 files in 24-48 hours for £295 + VAT, including content audit and validation.
Do small businesses need AI discovery files?
Small businesses arguably benefit more than large ones. When someone asks AI "recommend a local plumber in Leeds," larger competitors with established brand recognition might appear by default. AI discovery files level the playing field by giving small businesses the same structured visibility as enterprise competitors.
How can I check my current AI visibility?
Use our free AI Visibility Checker at ai-visibility.org.uk/ai-visibility-checker/ to analyse your website. It checks for all 10 AI discovery files, validates their content, and gives you a score out of 100 with specific recommendations. Takes 30 seconds, no signup needed.
How often should AI discovery files be updated?
Update them whenever core business information changes: new services, pricing updates, location changes, or product launches. Review quarterly to ensure accuracy. Any updates should be consistent across all 10 files to avoid contradictions.
Find Out Your AI Visibility Score
97% of UK businesses are invisible to ChatGPT. Our free checker tells you where you stand in 30 seconds.
Check Your ScorePublished: · Last reviewed: · Written by: Mark McNeece, Founder & Managing Director, 365i
Editorially reviewed by: Mark McNeece on · Our editorial standards
