← Jargon buster
The AI era

Agent readiness

Agent readiness is a score for how well AI systems can find, fetch and make sense of your website. It covers discoverability, crawler access, machine-readable content and agent capabilities, and it is the first thing to check when you are missing from AI answers.

Why it comes before content work

When a business is absent from AI answers, the instinct is to blame the writing and start rewriting pages. Often the cause is far more mundane: a line in a file that tells the relevant crawler to go away, a homepage that refuses non-browser requests, a site with no sitemap and nothing to follow.

Those are quick fixes with immediate effect, and it is worth ruling them out before commissioning content. Checking costs nothing, which is the point: the scan is ordinary web requests, so it can run daily without consuming anything.

Crawlers are split by purpose, and that distinction matters

This is where most scoring gets it wrong. AI crawlers do three different jobs and they deserve different treatment.

Search crawlers and user-fetch crawlers are how you get into answers. Blocking those removes you from AI results, and that is a genuine failure worth flagging loudly.

Training crawlers are the ones that collect material to train models. Blocking those is a legitimate editorial decision that plenty of publishers make deliberately. We report it so you know where you stand, and we never score it against you, because it is a choice and not a mistake.

Unmeasurable is not the same as failing

If a check cannot be completed, we mark it unknown and take it out of the scoring rather than counting it as a fail. A site that is temporarily unreachable should not come back reading as a site that fails every check.

The same applies to robots.txt: a file that times out or returns a server error is not the same as no file at all. Only a genuine 404 means "nothing is restricted here". Treating an unreadable file as permission is exactly the kind of assumption that produces confidently wrong findings.

It is a step function, so the history matters more than the score

Free tools that will hand you a readiness score are now everywhere, and they largely check the same handful of things. Today’s number is close to a commodity, and you can work most of it out by hand in twenty minutes.

The score sits flat for weeks and then jumps when somebody edits a file. So the line is not the interesting part: the interesting part is which checks changed status, and on what day. That is what turns "our AI visibility fell off at some point" into "this deployment, this check, this Tuesday", and it is the one thing a one-off scan can never tell you.

It is also why we store a scan per site per day rather than offering a button to press: the overall score, the four category scores, and every individual check with what was found. A fix that quietly gets undone six weeks later is indistinguishable from a fix that held, unless something was watching in between.

The full history is readable by your own AI assistant over the MCP connector, which is what makes progression reportable rather than merely visible: you can ask what changed since March and whether it lines up with a traffic drop.

Where this shows up in the tracker

Every term in this glossary is something the product actually measures or acts on. Plans start at £29 a month, with rank tracking, AI visibility, the daily AI readiness scan, security and domain health, Search Console, backlinks and audits on every one of them.