Free Agentic Browsing Audit
Is your website ready for AI agents?
An agentic browsing audit tests whether an AI agent can actually use your website — not just read it. Google added an Agentic Browsing category to Chrome Lighthouse that checks four things: your accessibility tree, cumulative layout shift, whether you have an llms.txt file, and WebMCP support. The tool below runs those same checks on your live page and tells you what an agent hits. Fix the accessibility tree first — it is what nearly every contractor site fails, and everything an agent does depends on it.
Straight answer up front: agent-readiness is not a confirmed ranking factor and there is no study showing it improves AI citations. This is positioning for when agents start outnumbering humans on your site — plus two of the four checks make your site better for real people today.
Can an AI agent actually use your website?
AI agents are starting to visit sites on a customer's behalf — comparing companies and requesting quotes. We run the same agentic browsing checks Google built into Chrome and show you what an agent hits when it lands on your site. No email to run it.
What does an agentic browsing audit actually check?
Four checks, in the order you should fix them. The first two are the ones that decide whether an agent completes a task on your site or gives up and goes to a competitor.
1. Accessibility tree
The stripped-down map of your page that agents navigate by — buttons, links, form fields and their names.
This is the one nearly every contractor site fails. A styled <div> that looks like a button has no name and no role, so an agent literally cannot see it as clickable. Fix this first: everything else an agent does depends on it.
Typical time to fix: Hours to days
2. Cumulative Layout Shift (CLS)
How much your page moves while it loads. Target is 0.1 or lower.
Agents move through a page far faster than a person. If your header drops down after a banner loads, the agent’s click lands on nothing, it reads that as an error, and it leaves. There is no second attempt.
Typical time to fix: Hours
3. llms.txt
A plain markdown file at yoursite.com/llms.txt that summarises your business and links your key pages.
Google recommends it, though John Mueller has confirmed Google Search itself does not consume it. It is roughly 20 minutes of work and there is no downside, so it is worth doing — just do not expect it to move rankings on its own.
Typical time to fix: About 20 minutes
4. WebMCP
An experimental protocol that tells agents exactly which actions they can take — submit a form, request a quote, book a job.
Genuinely experimental, and Lighthouse marks it "not applicable" on almost every site rather than failing it. Only worth building if you have real transactional flows and the technical capacity to secure them. For most contractors: watch, don’t build.
Typical time to fix: Hours to days
Why does agent-readiness matter for a contractor?
For twenty-five years a crawler visited your site, took a copy, and left. Agents behave differently: they act on a customer's behalf. A homeowner who needs an HVAC replacement can now tell an AI what they need and have it compare local companies, pull quotes, and hand back a shortlist — before the homeowner speaks to anyone.
That makes the shortlist the whole game. If an agent lands on your site and cannot find a name for your “Get a Quote” button, you are not on the shortlist, and nobody — including you — ever finds out why. Meanwhile every one of your competitors who happens to use a properly built form stays in.
Being found is a separate problem from being usable. If you also want to know whether AI names your business in the first place, that is what our AI search visibility checker answers. Agent-readiness decides what happens after the agent arrives.
Agentic browsing audit FAQs
What is an agentic browsing audit?
An agentic browsing audit checks whether an AI agent can navigate and use your website, rather than just read it. Google added an Agentic Browsing category to Chrome Lighthouse (version 13.3, Chrome 150+) that tests four things: your accessibility tree, cumulative layout shift, whether you have an llms.txt file, and WebMCP support. This free tool runs those same checks on your live page.
What does it mean for a website to be AI agent ready?
It means an AI agent acting for a customer can land on your site, understand its structure, identify your buttons and forms, click accurately, and complete a task like requesting a quote. Traditional crawlers only read and leave. Agents interact — so a site that reads fine for Googlebot can still be unusable for an agent.
Is being agent-ready a Google ranking factor?
No. Google has not confirmed agent-readiness as a ranking factor, and there is no published study showing it improves AI search citations either. It is positioning, not ranking. The value is being usable when agents start showing up in volume — and two of the four checks (accessibility tree, CLS) improve your site for real human visitors and screen-reader users today regardless.
Does Google use llms.txt?
Google Search does not. John Mueller has said Google does not use llms.txt and he is not aware of another crawler that does. At the same time, Chrome Lighthouse now checks for a valid llms.txt as part of its agentic browsing category. Both things are true: it is not a ranking signal, but it is a recommended agent-readiness signal that costs about 20 minutes.
Why did my site score 3 out of 3 but still show failures?
Lighthouse marks some checks "Not applicable" rather than failing them — WebMCP on virtually every site, and a missing llms.txt returns a 404 which Lighthouse treats as not applicable too. Those never count against your fraction. A perfect fraction means you passed everything that was scored, not that there is nothing left to do.
How do I fix a failing accessibility tree?
Give every interactive element a programmatic name and a real role. Replace styled <div> click handlers with actual <button> and <a> elements, label every form field, add accessible names to icon-only buttons, and remove aria-hidden from anything that is still focusable. Run the scan again after each pass, starting with your highest-value pages.
Want this fixed for you?
Accessibility tree and layout stability are core technical work — the same foundation that earns Google rankings and AI citations. It is already part of how we build. Tell us your site and we'll walk you through what we'd change and why.
Get a quote