Is your website ready for AI agents?

Your site has a
third audience.

Most sites turn them away at the front door. You never find out, because a page that fails an agent still looks perfectly fine to you.

Find out what yours does.

Free AI agent readiness audit. No email. Eight checks in 15 seconds.

What your site was built for

  1. Humans Your site: built layout · accessibility · brand
  2. Search engines Your site: maybe sitemaps · schema · meta tags
  3. AI agents Your site: nothing discovery · readability · interaction

How ready is the web?

Most sites are not ready.

We scan public URLs against the same deterministic rubric. Here is what we have found so far.

Sites scanned

3,567

Checks per scan

31

Average score

53/100

Most failed checks

  1. Agent discoverability 99%
  2. Agent readability 96%
  3. Agent interaction 94%

By outcome

Failing: 872 Partial: 1,996 Scoring 80+: 699

Does SEO cover this?

Google reads you. Agents might not.

Search crawlers render JavaScript. The crawlers behind ChatGPT, Claude and Perplexity do not. Your rankings were earned by a machine that reads your site differently.

42%

of JavaScript-rendered content never gets indexed by AI systems

Learn more at Onely, February 2026, opens in a new tab

#1

on Google and still invisible to ChatGPT, Claude and Perplexity

Learn more at Vercel and MERJ, opens in a new tab

57.5%

of all web traffic is now machines, not humans

Learn more at Cloudflare, June 2026, opens in a new tab

What do we check?

Agents ask four different questions.

Finding you, reading you, and acting on your site are three separate problems. A fourth watches what the engines say about you afterwards.

15Checks

Discovery

Can an agent find you and understand what you offer?

16Checks

Readability

Can an agent read your content without guessing?

13Checks

Interaction

Can an agent complete a task on your site?

6Checks

Representation

Can an agent describe you accurately, month after month?

What do we fix?

We hand you the fix,
with the finding.

Every failed check has a fix, and the crawl already gathered what the fix needs. So we fix it for you.

robots.txt Generated patch
- User-agent: * - Disallow: / + User-agent: GPTBot + Disallow: / # training + User-agent: ChatGPT-User + Allow: / # citation + Content-Signal: ai-train=no, search=yes

Deployable files

written from your crawl

Code patches

matched to your stack

Written instructions

where a human has to decide

What does it cost?

You just stop being remembered.

The agent reads what it can, answers with whatever it found, and moves on.

The answer goes to someone else

An agent recommends whoever it could read. There is no second page of results to be on.

Wrong answers, your name on them

Stale structured data and unreadable pages mean an agent quotes a price or a claim you retired years ago.

No way to catch up later

The months an agent skipped you are gone. Unlike ads, there is nothing to switch back on.

Isn't this just SEO?

Others tell you what is wrong.

Most tools watch whether AI mentions your brand. Nobody checks whether an agent can use your site once it arrives, and nobody hands you the fix.

Tool What it measures What it leaves you with
Lighthouse Four agentic checks, pass or fail A pass rate, no score, nothing to act on
Ahrefs Whether AI mentions your brand A visibility number, and no reason for it
Semrush Mentions and sentiment, on simulated prompts A dashboard your developers cannot use
Code4X Whether an agent can find, read and use your site A score, the evidence behind it, and the files that fix it

Where do we start?

Three ways in.

Every engagement starts the same way. A scan, then a conversation with an engineer.

One-off

Full audit

Every dimension scored, evidence for each lost point, the generated fix files, and an MCP connection to query it all.

from $4,500

Delivered in 5 working days

Most common

Retainer

Monthly re-audit on the same rubric, before and after deltas, and drift monitoring as the standards move.

from $1,800/mo

Rolling, three-month minimum

Premium

Live agent testing

A real agent driven through your key flows, with a record of where it succeeded and where it stalled.

from $6,000

Scoped per site

Start here

Free scan

Fifteen seconds, no email. You will know whether any of the above is worth a conversation.