AI agents are already visiting your website. Can they actually use it?
They fetch your HTML, parse it, and try to finish a task: find a price, compare a spec, start a checkout. We audit whether that works across 100 points of machine readability, then fix what doesn’t.
Free instant scan against the published 100-point rubric. No email, no sign-up — you get the report link in about a minute.
Illustrative, drawn from the pattern we see most often: respectable content, weak machine access.
Why this matters now
Your next customer may never see your website
They will ask an assistant, and the assistant will read your site on their behalf, or fail to and recommend a competitor it could read.
Rendering is the silent killer
Most agents fetch plain HTML and never run your JavaScript. If your content only appears after hydration, they see an empty shell, while your browser test looks perfect.
Your CDN may be blocking them
Cloudflare blocks AI crawlers by default, before robots.txt is ever read. A perfect robots.txt and a 403 at the edge is the single most common failure we find, and almost nobody knows it is happening.
Structure beats prose
An agent cannot infer that a div is a price. Schema, semantic landmarks and real form labels are what turn a page from readable into actionable.
The rubric
Four pillars, 25 points each
Every check is automatable from an unauthenticated crawl, and every result carries the evidence behind it.
- A
Discovery & access
Can agents find and reach the content at all?
- B
Understanding
Once fetched, can a machine parse what this site is and sells?
- C
Actionability
Can an agent actually do things — navigate, search, fill forms, transact?
- D
Trust & freshness
Should an agent — or the model behind it — trust and cite this site?
How it works
Three steps, no instrumentation on your side
We need nothing installed and no access to your analytics. Everything we measure is what a stranger's agent would see.
- 1
We crawl like an agent does
Your homepage plus up to five key pages, fetched twice, once as raw HTML and once fully rendered, and again under the GPTBot, ClaudeBot and PerplexityBot user agents. The gap between those fetches is usually where the problem lives.
- 2
You get a score and the evidence
100 points across four pillars, a letter grade, and every failed check backed by what we actually captured. No finding without evidence: if we cannot show you the response, we do not report it.
- 3
We fix what is worth fixing
Findings ranked by impact and effort, so the cheap high-value work is obvious. Then we implement it, in your codebase or your CMS.
Proof, not promises
We publish our own score
We sell agent readiness, so it would be absurd to hide behind a marketing site that fails its own audit. Ours is public, with the evidence.
See the leaderboard
Real sites ranked by what agents can actually do on them — ours included, measured by the same pipeline — with the full evidence one click away.
Browse the leaderboard →
Read the methodology
The complete 100-point rubric, published openly and versioned. Disagree with a check and you can tell us exactly which one.
Read the rubric →
Questions
Common questions
Is this just SEO with a new name?
No. SEO optimises for a search engine that renders pages and ranks links. Agent readiness optimises for a program that fetches your HTML, parses it, and tries to complete a task. A site can rank first on Google and still be unusable to an agent, most commonly because the content only exists after JavaScript runs, or because the CDN returns 403 to the agent’s user agent.
How is the score calculated?
Four pillars worth 25 points each. Every check is automatable from an unauthenticated crawl and returns the evidence we captured, not an opinion. The whole rubric is published openly, including how each check is scored and what would change it.
Do you crawl politely?
Yes. We obey robots.txt for our own user agent, fetch at most six pages plus six file probes per audit, hold one concurrent request per host, and time out at ten seconds a page. The crawler is documented on its own page, and you can opt out.
What does it cost?
Audits start at $2,000. Implementation engagements typically run $4,000–$25,000, scoped against your architecture rather than your page count. More on how we scope.
Who is behind this?
AgentFriendlyRank is a service of Socio360, a sole proprietorship registered in Karnataka, India. More about how we work.
Find out what an agent sees when it visits your site
A free audit against the full rubric, with the evidence. If there’s nothing worth fixing, we’ll tell you that too.