AI: 01

Agent Readiness Audit

Find out how your site scores.
Before someone else does.

You can’t fix what you haven’t measured.

Since May 2026, anyone can run Lighthouse against your website and see an Agentic Browsing result — a count of how many agent-readiness checks you pass. Your competitors can. Your prospects can. A journalist can. Here’s the thing almost nobody has noticed: that number is nearly useless as a measure of how ready you are. Lighthouse runs six checks in this category. Three of them test WebMCP, an experimental standard that requires a browser feature PageSpeed Insights doesn’t currently have — so they always return “not applicable”. A fourth checks whether your llms.txt file follows the recommended format, which is skipped entirely if you don’t have one. That leaves two checks that actually score: whether your accessibility tree is well-formed, and whether your layout is stable. So a competently built site gets 2 out of 2 and looks perfect — while its prices sit in images no machine can read, its sitemap is broken, and it has no crawler policy at all. We know, because that was our own site until we checked.

What we look at

The Lighthouse checks, properly interpreted. WebMCP support, agent-centric accessibility, layout stability, and file discoverability. We run them, then explain what each result actually means for your business rather than leaving you with a colour-coded chart.

Structured data and schema. Whether your organisation, products, services, opening hours, pricing and FAQs are marked up in a way machines can read without inference. This is the single highest-value area for most sites, and the most commonly neglected.

Content legibility. Pricing inside a JPEG. Key details locked in a PDF. Critical copy that only appears after a JavaScript render. Every one of these is invisible to a large slice of the systems now describing you.

Form and interaction structure. If an agent is going to complete an enquiry, booking or purchase on a customer’s behalf, it needs labelled fields, clear states and predictable flows. Most forms fail this. Yours might pass — accessible forms and agent-usable forms overlap heavily, so if you’ve done accessibility work you may be further ahead than you think.

Crawl and access policy. What your robots.txt currently says to AI crawlers, which of them you’re blocking deliberately, which you’re blocking by accident, and whether that’s the position you actually want to hold.

How you’re currently described. A snapshot of what the major assistants say about your business today, and where they’re getting it from.

What you get

A written report with a headline score, a ranked list of issues by impact against effort, and a roadmap you can hand to any competent developer. Plus a call to walk through it, because a PDF nobody reads has never improved anything.

How do I find out where I stand?

Agent readiness is assessed as part of our Technical Business Benchmark — a full diagnostic covering your competitive position, your site’s technical health, and whether AI can find, read and use you. £750 + VAT, credited in full against any work you commission afterwards.

Book your audit →

What happens next is up to you

Take the roadmap and run it yourself, or let us implement it through Supercharge Your Site. No pressure either way — the report is genuinely yours.

  • What is the Agentic Browsing score in Lighthouse?

    It’s a count of how many agent-readiness checks your page passes. Lighthouse runs six checks in this category, but three test WebMCP — an experimental standard PageSpeed Insights can’t currently evaluate — and a fourth is skipped if you have no llms.txt file. In practice most sites see a result out of two.

  • Is a good Agentic Browsing score enough?

    No. A competently built site scores full marks while its prices sit in images no machine can read, its sitemap is broken and it has no crawler policy at all. The score doesn’t test any of those things.

  • How do I find out if my site is agent ready?

    Agent readiness is assessed as part of our Technical Business Benchmark, which is £750 + VAT and credited in full against any work you commission afterwards.