13%

Last updated: March 2026

We checked 30 UK small and medium-sized business websites across nine industries to see whether they had an llms.txt file in place. The short answer: almost none of them did.

---

TL;DR

• Only 4 out of 30 UK SME websites we checked had an llms.txt file (13.3%)
• Marketing agencies had the highest adoption at 50% — still only half
• Law firms, estate agents, dentists, personal trainers, restaurants and retailers: 0%
• Accountancy firms showed early signs of adoption at 22%
• llms.txt is still an emerging convention with no formal standard, but adoption is beginning to appear in digitally aware sectors

---

What is llms.txt?

llms.txt is a plain text file placed at the root of a website (e.g. yoursite.co.uk/llms.txt) that provides a structured summary of what the site is about, who runs it, and what its key pages cover. It is designed to help large language models — the technology behind ChatGPT, Claude, Perplexity and others — understand a website more quickly and accurately.

The concept was proposed in late 2024 and has since gained traction among developers and some forward-thinking businesses. It is not yet a formal web standard, and there is no confirmed evidence from any AI platform that the presence of an llms.txt file directly influences whether a business gets cited or recommended. It remains an emerging convention.

That said, the logic behind it is straightforward: if AI systems are going to read your website, giving them a clear summary of what you do and how your content is structured may help reduce the chance of misinterpretation. Whether that translates into measurable visibility improvements is still an open question.

---

What we checked

We selected 30 UK SME websites at random across nine industries:

• Accountancy (9 sites)
• Marketing agencies (4 sites)
• Estate agents (4 sites)
• Law firms (3 sites)
• Plumbers (3 sites)
• Personal trainers (3 sites)
• Dentists (2 sites)
• Restaurants (1 site)
• Retail (1 site)

For each site, we checked whether an llms.txt file was accessible at the standard root location. We also checked several related signals including robots.txt configuration, schema markup, FAQ schema, trust pages (about, contact, privacy) and whether AI crawlers were being blocked.

This is not a statistically representative survey of all UK businesses. It is a snapshot across 30 real websites, chosen to reflect the kinds of SMEs that typically want to be found by potential customers online.

---

The findings

llms.txt adoption: 13.3%

Only 4 out of 30 websites had an llms.txt file. The vast majority of UK SMEs we checked have not implemented this file.

Industry	Sites checked	Has llms.txt	Adoption rate
Marketing Agency	4	2	50%
Accountancy	9	2	22%
Legal	3	0	0%
Estate Agent	4	0	0%
Plumber	3	0	0%
Personal Trainer	3	0	0%
Dentist	2	0	0%
Restaurant	1	0	0%
Retail	1	0	0%

Marketing agencies had the highest adoption, which is perhaps unsurprising — these are businesses whose work involves staying close to emerging digital trends. Even so, half of the marketing agencies we checked did not have the file.

Accountancy firms showed early adoption at 22%, suggesting that some firms in this sector are beginning to consider how AI platforms interpret their online presence.

Every other industry we checked — law, property, trades, health, fitness, hospitality and retail — showed zero adoption.

---

Related findings from the same check

While checking for llms.txt, we also gathered data on several other AI visibility signals. These paint a broader picture of how prepared UK SME websites are for an AI-driven search environment.

23% are blocking AI crawlers

7 out of 30 websites had robots.txt rules that block one or more AI crawlers.

AI Crawler	Sites blocking	Percentage
GPTBot (ChatGPT)	7	23.3%
ClaudeBot (Anthropic)	5	16.7%
PerplexityBot	1	3.3%
Any AI crawler	7	23.3%

It is worth noting that blocking AI crawlers is a legitimate choice — some businesses may prefer not to have their content used by AI platforms. However, in several cases, these blocks appeared alongside other signals suggesting the business does want to be discovered online (active SEO, FAQ content, schema markup). This raises the possibility that some businesses may be blocking AI crawlers without fully realising the implications, particularly where the robots.txt was configured by a developer or plugin rather than as a deliberate business decision.

Only 4% have Person schema

Just 1 out of 25 accessible websites had Person schema markup connecting a named individual to the business entity. This means 96% of the websites we checked have no structured way for AI systems to understand who runs the business, what their expertise is, or how they relate to the organisation.

Whether Person schema directly influences AI recommendations is not established. But AI platforms do appear to value entity relationships — understanding that a specific person founded, leads or is the primary expert behind a business. Without structured markup, that connection relies entirely on AI systems inferring it from unstructured text, which is inherently less reliable.

52% have FAQ content but no FAQ schema

This was one of the most striking findings. 13 out of 25 websites had question-style content on their homepage — headings structured as questions like "How much does it cost?" or "What services do you offer?" — but did not have FAQ schema markup (FAQPage structured data) to help AI systems identify and extract those answers.

The content exists. The structured signal that tells AI platforms "this is a question and answer" does not.

Again, we cannot say with certainty that FAQ schema directly improves AI citation rates. Research from third parties suggests that structured content with clear question-and-answer formatting may perform better in AI-generated responses, but this is an evolving area. What we can say is that more than half the sites we checked have done the hard work of creating FAQ content but have not taken the additional step of marking it up in a way that AI systems can easily parse.

Only 36% have all three basic trust pages

We checked for three foundational trust pages: About, Contact and Privacy Policy. Only 9 out of 25 accessible sites (36%) had all three.

Trust page	Sites with page	Percentage
Privacy Policy	21	84%
Contact page	14	56%
About page	13	52%
Terms page	7	28%
All three (About + Contact + Privacy)	9	36%

Privacy policies were the most common, likely driven by GDPR requirements. But about pages and contact pages — the pages that help establish who a business is and how to reach them — were missing from roughly half the sites we checked.

These pages serve multiple purposes beyond AI visibility. They build trust with human visitors, satisfy regulatory requirements, and provide the kind of entity information that any system — human or machine — uses to assess whether a business is legitimate and established.

---

What this might mean

We want to be clear about what this data shows and what it does not.

What it shows: The vast majority of UK SME websites we checked have not implemented llms.txt, and many are missing other signals that could help AI systems understand their business. This is a snapshot of current adoption, not a prescription.

What it does not show: We cannot confirm that having an llms.txt file, FAQ schema, Person schema or any individual signal directly causes improved visibility in AI search results. The relationship between specific technical signals and AI citation behaviour is still being studied across the industry.

What seems reasonable to observe: AI platforms are increasingly being used by consumers and business buyers to discover and evaluate companies. These platforms need to understand what a business does, who runs it, and whether it is credible. The signals we checked — llms.txt, schema markup, trust pages, crawler access — are all ways a business can make that information clearer and more accessible. Whether AI platforms currently weight these signals heavily is uncertain. Whether they will in future seems more likely than not.

The businesses that are already implementing these signals are, at minimum, ensuring their information is available in a structured format should AI platforms choose to use it. The businesses that are not may find themselves harder for AI systems to interpret accurately — not because they are being penalised, but simply because the information is not there in a form that machines can easily process.

---

Industry breakdown

Legal (0% llms.txt, 0% schema): Law firms were the least prepared across almost every signal we checked. None of the three firms had any schema markup, llms.txt, or FAQ content. This is notable given that legal services involve high trust requirements and consumers increasingly use AI to research solicitors and legal options.

Marketing agencies (50% llms.txt, 100% schema): The most prepared sector, though still with significant gaps. Marketing agencies are closest to the technology and most likely to have heard of llms.txt, which explains the higher adoption. Even so, Person schema was absent across all four.

Accountancy (22% llms.txt, 100% schema): Showing early signals of awareness. Two of nine firms had llms.txt, and all had some form of schema markup. However, only one had FAQ schema despite several having FAQ-style content.

Estate agents (0% llms.txt, 50% schema): Mixed picture. Half had basic schema but none had llms.txt or FAQ schema. One was actively blocking all three major AI crawlers.

Trades, health, fitness, hospitality (0% llms.txt across all): These industries showed no llms.txt adoption. Most had basic schema markup (likely from their website platform rather than deliberate implementation) but lacked the more specific signals like FAQ schema, Person schema or sameAs links.

---

Methodology

• Sample: 30 UK SME websites selected from web search results across nine industries
• Date: 12 March 2026
• Checks performed: llms.txt presence, robots.txt AI crawler blocks (GPTBot, ClaudeBot, PerplexityBot), schema markup (any, Organization, Person, FAQ, sameAs), FAQ content detection via question-style headings, trust page presence (about, contact, privacy, terms), homepage alt text coverage
• Limitations: This is a convenience sample, not a statistically representative survey. Results reflect what we found across these specific 30 websites and should not be extrapolated to all UK businesses. Homepage-only checks may miss signals present on internal pages. Schema detection was based on homepage HTML source and may not capture dynamically loaded structured data.

---

FAQ

---

What is llms.txt?

llms.txt is a plain text file placed at the root of a website that provides a structured summary for large language models. It typically includes a description of the business, its key services, and links to important pages. The format was proposed in 2024 and is gaining adoption but is not yet a formal web standard.

Does having llms.txt help my business appear in AI search?

There is no confirmed evidence that llms.txt directly influences AI citation decisions. It is an emerging convention that may help AI systems understand your website more accurately, but its impact on visibility has not been independently verified.

How do I create an llms.txt file?

You can create one manually by writing a structured text file and uploading it to your website root. Rank4AI offers a free llms.txt creator tool at rank4ai.co.uk/tools/free-llms-txt-creator that generates the file based on your website content.

Why are some businesses blocking AI crawlers?

Robots.txt rules blocking AI crawlers like GPTBot or ClaudeBot may have been set deliberately (some businesses prefer not to have their content processed by AI) or may have been configured by a developer or website plugin without the business owner's knowledge. If you want AI platforms to be able to read your website, it is worth checking your robots.txt file.

What is Person schema and why does it matter?

Person schema is structured data markup that identifies a named individual and their relationship to a business or organisation. It helps AI systems understand who is behind a company, what their expertise is, and how they connect to the business entity. In our check, only 1 out of 25 websites had this markup.

What is FAQ schema?

FAQ schema (FAQPage structured data) is markup that identifies question-and-answer content on a page. It helps search engines and AI systems recognise that specific content is structured as questions with corresponding answers, making it easier to extract for use in generated responses.

Does schema markup help with AI visibility?

Schema markup helps AI systems understand the structure and meaning of your content. Whether it directly influences AI citation decisions is still being studied. What is clear is that structured data reduces ambiguity — it helps machines understand what your business is, what it does, and how its content is organised.

How many UK businesses have an llms.txt file?

Based on our check of 30 UK SME websites in March 2026, 13.3% had an llms.txt file. Adoption was highest among marketing agencies (50%) and accountancy firms (22%), with zero adoption across law, property, trades, health, fitness, hospitality and retail.

Should I implement llms.txt on my website?

The file is quick to create and there are no known downsides to having one. If you want AI platforms to have a clear, structured understanding of your business, it is a reasonable step to take. Whether it will directly improve your AI visibility is not yet proven, but it ensures the information is there if AI systems choose to use it.

What other signals should I consider?

Based on our research, the signals most commonly missing from UK SME websites were: Person schema (96% missing), FAQ schema (88% missing), llms.txt (87% missing), About page (48% missing), Contact page (44% missing) and Terms page (72% missing). Addressing any of these gaps ensures your business information is more accessible to both human visitors and AI systems.

---

This research was conducted by Rank4AI as part of our ongoing work understanding how UK businesses appear — or don't appear — in AI-powered search platforms. We publish original data alongside curated industry statistics to help UK businesses make informed decisions about AI search visibility. Our findings are observational and should not be taken as guarantees of specific outcomes.

For more on how AI platforms interpret business websites, see our guides on AI search visibility, the Five Signal framework, and our free llms.txt creator tool.

Want to understand how these trends affect your business?

Start Your AI Visibility Review