What is the difference between the free scan and the £99 report?

The free scan looks at your homepage and up to two key pages, giving you your overall glassmark score, the six category scores, and your top three priorities. It is enough to show you where the gaps are and whether you want to go deeper. The £99 report scans up to 26 pages across your site and gives you every finding across all six categories, with developer-ready code snippets pre-filled with your actual business details. You also get a tailored prompt kit to test your visibility in real AI tools, a downloadable PDF you can share with a developer, and five rescans across 90 days so you can measure your progress as you make fixes.

Who is glassmark for?

glassmark is built for businesses in the visitor economy: places to stay, places to eat, attractions, activities, and the suppliers behind them. If someone can book a room, reserve a table, buy a ticket, or book an experience with you, glassmark is built for your business. It is most useful if you have noticed your website traffic or enquiries softening and are not sure why, or if you have heard AI is changing how people find businesses and want to know where you stand. You do not need to be technical. The scan tells you what is wrong in plain language, and the report is written so you can act on it yourself or hand it to whoever looks after your website.

What do I actually get from a scan?

The free scan gives you your glassmark score, six category scores, and your top three priorities, enough to see where the gaps are. The £99 report gives you every finding across all six categories, with developer-ready guidance for each one, a downloadable PDF you can hand straight to a web person, a prompt kit to test your visibility in real AI tools, and five rescans across ninety days so you can watch your score move as you make fixes. In short: the free scan tells you there is a problem. The report tells you exactly what to do about it, in priority order.

What happens if I use all my scans?

The £99 report includes six scans in total, your first scan plus five rescans, across a ninety-day window. That is enough for most businesses to make their fixes and measure the improvement. If you use all six and want to keep going, you can top up with more scans rather than buying a whole new report. Top-up options are on the pricing page.

Is this a subscription?

No. Both the £99 report and Work with Wendy are one-off purchases. There's no recurring billing, no auto-renewal, and no subscription trap. The £99 gives you six scans across ninety days because that's a realistic implementation window, and after that you simply have the results you already have. Nothing auto-charges. Here's the caveat worth knowing though. AI readiness is a moving target. The models change, the standards evolve, and the signals that matter most today may shift in twelve or eighteen months. glassmark evolves alongside that. So if you run a scan today, act on the findings, and then decide in six months that you'd like to check in again because the landscape has moved, that's a fresh decision you make when the time feels right. Not automatic, not assumed. We think it's more upfront to say 'this is one moment in time' than to pretend today's fixes are forever. The way AI reads websites today isn't the way it will read them in two years, and we'd rather be straightforward about that than pretend otherwise.

Do I need a developer to make the fixes?

Most fixes are manageable for anyone comfortable editing their website, though having a developer or web person involved usually makes the process smoother, especially for anything involving structured data. If you already work with someone who looks after your website, they'll be the right person to share the report with. If you don't have a developer, the Work with Wendy tier exists specifically to help with implementation.

How much work is this for my developer?

The answer is that it depends on your specific site, and the best person to tell you is whoever looks after it. Most of the fixes are configuration work rather than rebuilding anything, so they sit comfortably inside what most web people do day to day. The best approach is to start with your top three priorities, share the report with your developer, and have a conversation about what's involved for your specific setup. The £99 report is written so it can be handed straight to them as a brief, and the rescans let you measure progress as fixes land.

Will my booking system still work?

Yes. glassmark doesn't change anything about how your site takes bookings. What it checks is whether the key booking information, like prices and availability, is readable by AI agents. Most reservation widgets, across every provider, render inside JavaScript and are opaque to AI agents by design. The fix is to expose the key booking information at the page level using schema, so AI agents can read what they need without touching the widget itself. Your booking system continues to do what it does.

What if I use Wix, Squarespace, or WordPress?

glassmark works for sites built on any platform, because it's measuring the output that machines see, not the tool you used to build it. Most of the fixes in the report can be applied to any major CMS. WordPress has plugins. Squarespace and Wix have code injection features. The report explains your options based on the platform glassmark detects.

Is this only for UK businesses?

No. glassmark is built for any business in the visitor economy anywhere in the world. The scoring logic, the schema checks, and the AI readiness principles are all based on international standards like schema.org. The early customer base is UK-focused because that's where Wendy's network is, but the product has no UK bias in its scoring or its advice.

Do AI agents actually drive bookings yet?

The numbers are small today. Around 8% of UK holidaymakers used AI to plan or book their 2025 trips, up from almost nothing the year before. Hospitality websites lost an average of 6.7% of their organic traffic in late 2025 as Google started showing AI-generated summaries instead of links. The trajectory is what matters. AI-assisted discovery is growing quickly, and the operators who structure their sites for it now will be the ones AI agents recommend as the volume grows. glassmark exists because this is the moment to act, before the space gets crowded.

How accurate is the scan?

The scan reads your homepage plus two priority pages, checks for every schema.org type that matters for your business category, validates your structured data, examines your robots.txt and sitemap, and checks for llms.txt. It's a real diagnostic, not a marketing tool. The approach behind it is worth knowing. glassmark isn't a generic SEO audit with an AI label on the front. It's built by someone whose day job is building products with AI, which means the scanner is informed by what modern AI agents actually parse today, not by old assumptions about how search worked five years ago. Where it has to make an inference, for example detecting your primary location from page content rather than structured data, it tells you, because guessing from content is exactly the problem AI agents face too. One more thing worth flagging. glassmark went live in April 2026. Between now and mid-summer, the scoring rules, the entity detection, and the narrative quality are all being actively tuned against real scans of real businesses. The fundamentals are solid, but the edges are still being sharpened. If you run a scan and something looks wrong or misleading, email wendy@fourteenseed.com directly. Early feedback shapes the product, and early users are the people who help make it better for everyone else.

Will this affect my Google SEO?

The changes glassmark recommends usually improve your Google SEO as a side effect. Structured data, clean meta descriptions, and well-organised content help every kind of reader, human or machine. Google is increasingly reading the same structural signals AI agents rely on, so anything that makes your site clearer for ChatGPT or Claude tends to help your search rankings too. glassmark does not replace your SEO work. It sits alongside it, focused on the newer question of whether AI agents can read and recommend your business.

What counts as a visitor economy business?

Any business where someone can book a stay, reserve a table, buy a ticket, or book an activity. That includes holiday cottages, hotels, campsites, holiday parks, restaurants, pubs, cafes, attractions, museums, tour operators, activity providers, event venues, and destination marketing organisations. If you're not sure whether glassmark applies to your business, run the free scan and see.

Can I get a refund if I buy the report?

Yes. If the £99 report doesn't give you actionable, specific information about how to improve your site's AI readiness, email wendy@fourteenseed.com within fourteen days of purchase and you'll get a full refund. The guarantee exists because the product should earn its price on its own merits.

What happens to my data?

glassmark only reads pages that are publicly accessible on your website, the same pages any visitor or search engine can see. We do not access anything behind logins, passwords, or booking systems. We do not collect personal data about your customers. We do store the results of your scan so you can return to your report using your access code. Aggregate findings across all scanned sites help us build a benchmark of AI readiness in the visitor economy, and may inform reports for tourism bodies and industry publications. Your business is never identifiable in that aggregate data unless you have explicitly agreed to a case study.

Wendy Harris has spent two decades watching the internet evolve and fifteen years in tourism tech. She's built websites and tools for the visitor economy across the UK. She lives in Cornwall. glassmark is the AI readiness audit she built because the gap between what operators' websites say and what machines can read is widening fast, and no one was measuring it specifically for the visitor economy. You can read more on the about page or find her on LinkedIn.

How quickly will Wendy respond?

I respond to messages between 9am and 5pm Monday to Friday. If you message at the weekend or in the evening, you'll hear back the next working day. The form on the contact page goes straight to me.

When you buy the £99 report you get six scans in total, your initial scan plus five rescans, across a ninety-day window. Every time you make a fix you can rescan and see your score move. Rescans are triggered from the same results page using the unique access code that comes with your report.

Back to all posts

Is your website accidentally blocking AI?

By Wendy Harris·31 October 2026·5 minute read

Almost nobody decides to block AI. The setting arrived with a plugin, a security package, or a host's default, and nothing about the website looks any different afterwards. The site loads, the photographs are lovely, the booking button works. The only change is that a category of visitor can no longer read the page, and that visitor does not send a message to say so.

It is the quietest failure in this whole subject, because everything else keeps working.

Three kinds of AI visit, and only one of them is training

These are worth separating, because you might reasonably want a different answer for each.

The first is search-style crawling. OpenAI's OAI-SearchBot exists to surface websites in ChatGPT's search features, and Anthropic's Claude-SearchBot navigates the web to improve the quality of search results. These are the crawlers behind AI answers that cite their sources. If you want your business to come up when someone asks an assistant where to stay in your town, these are the ones that matter most.

The second is training. GPTBot collects content that may be used to train OpenAI's models, and ClaudeBot does the same for Anthropic. Blocking these is a perfectly reasonable position, and plenty of businesses hold it deliberately.

The third happens when a person asks about you directly. ChatGPT-User and Claude-User fetch a page because someone requested it in that moment. That is the closest thing to a customer typing your name into a search box.

The three usually get treated as one thing. A single line in a file, or one tick box in a security dashboard, can switch off all of them at once, including the ones you would have wanted.

The date worth knowing, and what it actually covers

On 1 July 2026, Cloudflare introduced three bot categories, Search, Agent, and Training. From 15 September 2026, for domains newly onboarding to Cloudflare, Training and Agent crawlers will be blocked by default on pages that display advertisements, while Search remains allowed by default.

Read the scope rather than the headline. It applies to new domains, not existing ones. It applies to pages carrying advertising, not to every page. Search is unaffected. Most visitor economy websites carry no advertising and are not newly onboarding, so most will not be touched by this at all.

It is still worth knowing, for two reasons. It is the clearest sign yet that access is becoming something decided by infrastructure defaults rather than by you. And it makes concrete a question that has been vague for two years: do you actually know what your site allows today?

Where accidental blocking comes from

Almost never from a decision. Usually from one of these four.

A robots.txt file written by a plugin or a theme. Plenty of SEO and security plugins now offer AI crawler blocking, some of them suggesting it during setup. It gets agreed to once and never revisited.

A bot protection setting at the host or the CDN. These are described in terms of scrapers and bad bots, and the category names rarely map onto the three kinds of visit above. Operators end up agreeing to something broader than they understood.

A holding page, staging password, or geographic restriction left in place after a rebuild. That blocks everyone equally, customers included, and it survives far longer than anybody expects.

Content that only appears after the page has loaded. Prices, opening hours, and menus often live inside a widget that fills in a moment later. It is not blocking, but the effect can be the same: whatever reads the page early finds the interesting parts empty.

The ten-minute check

Put your domain followed by /robots.txt into a browser, so yourbusiness.co.uk/robots.txt, and read what comes back. You are looking for two things: a blanket Disallow under User-agent: *, and any of the crawler names above listed individually. If GPTBot, ClaudeBot, OAI-SearchBot, or Claude-SearchBot appear in there, somebody made a decision on your behalf. Check that it is the decision you would have made.

Then ask whoever runs your hosting or your CDN a direct question: is bot protection switched on, and does it include AI crawlers? The phrasing matters, because "we block bad bots" is not an answer. What you want to know is which categories are blocked and who chose them.

Last, open one page, view the page source, and search it for a price or an opening time you know is on that page. If the text is there in the raw source, good. If it is not, that content is being added after the page arrives, and you should not assume every system waits for it.

That is the entire check. It costs nothing, it needs no developer, and it settles a question most businesses have never thought to ask.

A note on llms.txt

You may have been told to add an llms.txt file, a plain text summary of your site written for AI systems. Adoption has grown fast. Originality.ai tracked more than three million websites and found llms.txt instances rising 8.8 times in a year, reaching 36,120 by May 2026.

The usage evidence does not match the enthusiasm. Ahrefs looked at server logs across 137,000 domains and found that 97 percent of llms.txt files received no requests at all during May 2026. The largest single source of the requests that did arrive was SEO audit tools, at 21.7 percent.

Add one if you want to; it costs very little. Just do not file it under access solved. A summary nobody fetches does nothing for a site whose robots.txt is closed, and the file is no substitute for knowing what your own settings permit.

Access comes first

Every other kind of readiness work assumes something can reach the page. Clear pricing, structured information about your business, answerable opening hours, none of it counts for anything behind a closed door. That is why access is one of the categories in every glassmark scan, and why a Missing there matters more than a Missing anywhere else.

The good news is that this is the cheapest problem in the subject to fix. Unblocking a crawler is usually one line, one setting, and ten minutes of somebody's afternoon. It just has to be somebody who knows to look.

Read more

What does AI find when it reads your website? Free scan at glassmark.ai, no developer required.

Run your free scan

Found this useful?