When your customer asks AI, is your business the answer?
People no longer only type into Google. They ask ChatGPT, Gemini and Google’s own AI to find them a supplier, the right product, the best price. AI readiness is the work of making sure your website can be read, trusted and, increasingly, used by AI. Most sites can’t do any of the three yet.
Your AI-readiness score is part of the free audit. We run Google’s own test on your site, live, and you see the result before you spend a rand.
The engines your buyers ask. Every answer should orbit you.
What you get
A readiness score you can verify yourself
We run Google’s PageSpeed test on your site together, and read the Agentic Browsing line next to the four classic scores. It is the same test anyone can run. No mystery report: you watch the numbers appear.
A site AI can actually read
The crawlers behind ChatGPT, Claude and Perplexity do not run JavaScript. A site that builds its pages in the browser, which is how many modern sites work, hands them a blank page. Your content gets served ready-rendered, so every AI reads the same page your customers see.
Structured data that tells AI what you are
Schema is the machine-readable layer under the page. It is what AI systems use to ground their answers: who you are, what you sell, where you operate, what customers say. It gets written from your real business facts, nothing invented.
The gatekeepers, checked
Around a fifth of the web sits behind a firewall that blocks AI crawlers by default, and many owners have no idea. We check who is allowed in, let the right bots through, and register your pages with the index that feeds ChatGPT and Copilot.
llms.txt, done to the published spec
The plain-text index AI agents read first. Google’s test checks the format, not just that the file exists. Ours are generated from the site itself, so they never drift out of date.
Agent interaction: the site AI can use
The frontier work: your enquiry forms, bookings or products declared as actions an AI agent can complete for its user. In the US, ChatGPT already completes purchases at over a million stores. The browser side is in live trial in Chrome now. Being early here is the whole point.
Google already scores it. Most sites are failing it.
PageSpeed Insights, Google’s own testing tool, now grades every website on Agentic Browsing: can an AI agent read this page, trust it, and act on it? It sits right next to the Performance and SEO scores. Six checks, in two groups.
The accessibility tree is well-formed
An AI agent does not look at your design. It reads the structure underneath, and if buttons and fields have no proper names, the agent is guessing.
The layout holds still
Pages that jump around while loading break an agent mid-task the same way they annoy a person mid-click. Google measures the shift.
llms.txt follows the spec
A plain-text index of your site, written for AI to read. Google checks that it exists and that it follows the published format. Most sites have neither.
WebMCP tools registered
Your site tells the agent what it can do here (request a quote, book, buy) instead of leaving it to guess from the pixels.
Form coverage
Your enquiry and checkout forms declared as actions an agent can complete on a customer’s behalf.
Valid schemas
Each declared action carries exact, machine-checkable instructions, so the agent gets it right the first time.
These three run on WebMCP, a browser standard from Google and Microsoft that is in live trial in Chrome now. Early, deliberately: the sites that are ready when it switches on are the ones agents will use.
Don’t take my word for it. Put riaanmiles.com into PageSpeed and look at the Agentic Browsing line. Then put in your own site.
How it works
The readiness check
Part of the free audit. We run Google’s test on your site, live on the call, and go through what passes, what fails and what it means for how AI sees your business.
The findings, in plain English
A written rundown of what failed and why it matters: which fixes touch content, which touch the engine underneath, and what each involves. You decide with the evidence in front of you.
The work, logged
Rendering, schema, crawler access, llms.txt, agent tools, all done in the right order, every change on the record with the date and the reason.
Proof, not promises
We re-run the same public test when the work is done. Before and after, side by side, from Google’s own tool rather than my slides.
This site passes Google’s agentic-browsing checks today, with perfect accessibility, best-practice and SEO scores beside them. The test is public. Run it on this site, then run it on yours.
See the results →Common questions
01What is AI readiness, in one paragraph?
It is making your website work for a new kind of visitor: AI systems reading on behalf of your customers. That splits into three jobs. Can AI read your site (rendering, structure)? Does AI trust and understand it (structured data, crawler access, llms.txt)? And can AI act on it (forms and products declared as actions an agent can use)? Google now scores all of this in PageSpeed under Agentic Browsing.
02Is this just SEO with a new name?
No. It sits on top of SEO. Classic SEO earns you a ranking when a person searches. A growing share of buying questions now end inside an AI answer instead, and AI readiness makes sure your business can be read, quoted and chosen when they do. The two share foundations, which is why I do them together, but each has work the other never touches.
03My site is modern. Surely it is fine?
Modern is often the problem. Many recent sites build their pages in the browser with JavaScript, and the crawlers AI assistants send do not run JavaScript, so they get a near-blank page. Add the firewalls that block AI crawlers by default and a missing llms.txt, and a beautiful new site can be invisible to AI. The test settles it either way in about a minute.
04What is “agentic browsing”?
An agent is an AI that does things instead of just answering: “find me three quotes for X and fill in the enquiry forms.” Agentic browsing is that agent using websites. Google’s PageSpeed tool now scores how well a site supports this: three checks every site is graded on today, and three more for sites that declare actions agents can perform. That second group runs on WebMCP, a browser standard from Google and Microsoft currently in live trial in Chrome.
05Can AI really send you a lead or buy from my site?
Buying through AI is already real: in the US, ChatGPT completes purchases at over a million connected stores, and Google has launched agent checkout in its AI results. Agents filling in enquiry forms is the next step, and the browser standard for it is in live trial now. I will not promise you a flood of AI-driven leads next month. I will make sure that when agents arrive at your door, yours is a door that opens.
06What does it cost?
It depends entirely on where your site stands. Some sites need a handful of engine fixes, some need the rendering problem solved properly, which is bigger work. The free audit establishes which one you are, and the fee is quoted from the findings. You will know the number before any work starts.
Find out what AI sees when it reads your site.
The free audit now includes your AI-readiness score, run live on Google’s own public test and explained in plain English.
Get your free audit