All resources
Agency workflow·Beginner·6 min

Audit a client's AI readiness in 10 minutes

Before you promise a client more AI visibility, you need to know what's holding them back. The GEO Audit turns a site into a 0–100 readiness score with a prioritised fix list in a single pass — and now carries a separate Agent Readiness sub-score for the emerging class of AI agents. Here's how to run one and turn it into a deliverable.

What you'll be able to do: A prioritised, plain-language AI-readiness report for any client site — with fixes you can hand to a developer.

Transcript

Ask ChatGPT or Perplexity to recommend a business like your client's, and either it names them — or it doesn't. This is the ten-minute audit that tells you which, and exactly why. Most of the answer arrives in the first twenty seconds.

Adding a client takes about a minute — a name, and the one field that really matters: the website. That's the site the audit will read the same limited way an AI crawler does.

You don't need a full scan to check readiness, so skip that and open the GEO Audit. Hit Run audit, and while it works it fetches the homepage, the robots file, the sitemap and the llms.txt over plain HTTP with no JavaScript — the same narrow, stripped-back view an answer engine gets when it decides whether to cite you.

There's the score — a 77, rated Good — and the two numbers beside it tell the real story. Agent Readiness is a perfect hundred: live AI agents can fetch and read this site just fine. But Entity Authority is zero — technically ready, and yet the AI has no idea who this brand actually is.

In the insights, the word 'website' appears twelve times on this short page — the audit flags that as keyword stuffing, which the peer-reviewed GEO research calls the single worst tactic — and there isn't one cited statistic, quote, or link to an authoritative source. Real barriers, each with the citation lift it's costing you.

The Priority Fixes are a ranked to-do list, biggest win first. The top two together are worth eight points: anchor the brand with an Organization schema, and publish an XML sitemap. Click 'Draft the fix with Crunch' and it writes the exact code to paste, ready to hand to a developer.

Fix a few, re-audit, and watch the score climb — that's your proof the work landed, not a promise. Then drop the score, the fixes and the drafted code straight into a client report. A specific, credible action plan, in about ten minutes, on any domain you point it at.

  1. 1

    Run the GEO Audit

    Open the client's AI Visibility project, go to GEO Audit, and the domain is pre-filled from their saved website. Click Run audit — CrunchJunkie fetches robots.txt, the homepage, llms.txt and the sitemap the same way an AI crawler does (plain HTTP, no JavaScript) and scores concrete signals into one number.

  2. 2

    Read the score and categories

    The composite 0–100 score weights five categories: AI crawler access, content accessibility / SSR, structured data, technical hygiene and llms.txt. A high score means the site is well-positioned to be cited — it's a readiness measure, not a promise of mentions, which you track separately on the Visibility tab.

  3. 3

    Work the prioritised fixes

    The audit ranks every non-passing check by the points it would recover — highest-impact first. Each comes with a plain-language explanation, and "Draft the fix with Crunch" writes a concrete, copy-pasteable remediation grounded in that specific finding, ready to drop into a ticket.

  4. 4

    Check the Agent Readiness sub-score

    Alongside the GEO score is a separate Agent Readiness sub-score: can real-time AI agents fetch the site, read it without JavaScript, and parse a breadth of structured data. It's deliberately kept out of the main score because the category is still emerging — but it's a forward-looking differentiator few audits offer.

  5. 5

    Turn it into a client deliverable

    The audit history saves every run, so you can re-audit after the client ships fixes and show the score move over time. Drop the score, the prioritised fixes and the Draft-the-fix remediations into a report or a short email — a credible, specific action plan rather than vague advice.

Frequently asked

It fetches the site as an AI crawler does and grades: whether robots.txt and the CDN allow the AI crawlers, whether key content is server-rendered (not JavaScript-only), whether valid schema.org structured data is present, technical hygiene (sitemap, canonical, titles, HTTPS), and llms.txt. Present-but-incomplete schema is flagged with the exact missing properties.

The GEO score is about being crawled and cited by AI answer engines. The Agent Readiness sub-score is about the newer class of real-time AI agents that fetch on a user's behalf — whether they can reach the site, read it without JavaScript and parse its structured data. It's shown separately and doesn't affect the GEO score.

Yes. Every run is saved to the audit history with its date and score, so after a client applies fixes you re-audit and the history strip shows the improvement — a clean before-and-after for your report.

No, and we're deliberate about that. The score measures readiness — that the technical barriers to being cited are removed. Whether the brand is actually mentioned is tracked separately by Visibility % on the AI Visibility tab. Readiness is necessary but not sufficient; pair the two for the full picture.
GEO Audit docs

Try it on your own brand

AI search visibility and white-label reporting in one console — put this playbook to work inside the 14-day free trial.

Start free