Before you start tracking how ChatGPT or Perplexity talks about a brand, you need to know whether AI systems can actually read the site. That's what the technical audit is for.
If you haven't set up your brand yet, start with How to create your brand on AEO Copilot first.
How to run the audit
Once you're logged in, go to the Technical section at aeo-copilot.com/technical.
- Go to technical tab, on the left side of your screen
- Run the scan. It takes 30–60 seconds.
- You can also import a list of URLs, if your website is too big.
- Get your score and see what needs to be ready for optimization.
You'll get a score out of 100 and a breakdown by category.
!The Technical page showing a score of A/90, with 100% on Structure, Crawlability, and Authority
What the audit checks
Structure
Is the site's hierarchy clear? Are internal links logical? AI systems don't browse — they map. A site with scattered, unlinked pages is harder for a model to interpret than one with a clear structure. If this fails, check whether your main pages are actually linked from the homepage or navigation.
Crawlability
Can bots reach the site? This checks robots.txt, sitemap.xml, and whether pages are set to noindex. If crawlability fails, robots.txt is usually too restrictive or there's no sitemap.
Bot accessibility
Specific to AI bots — GPTBot, ClaudeBot, PerplexityBot. Some sites block them by accident, often through robots.txt rules set up for SEO purposes that were never revisited.
Schema markup
Structured data tells machines what a page is about. Organization schema on the homepage. FAQPage schema on FAQ sections. Article schema on blog posts. Without it, AI systems guess. Most sites just haven't added it yet — it's a missing fix, not a broken one.
LLMs.txt
A proposed file at the root of the site that summarises what the site contains for AI tools. It takes about 10 minutes to add, but treat it as optional. Google says its Search features, including AI Overviews, do not use it, and no major AI engine has confirmed that it affects citations. The audit checks for it because some AI tools and agents do read it.
Authority signals
HTTPS, proper meta tags, canonical URLs. If these fail, fix them first. They affect both SEO and AI visibility.
Readability
Is the content accessible without JavaScript rendering? Is heading structure logical — H1, then H2, then H3? JavaScript-heavy sites sometimes render nothing to a crawler. If your site is fully JS-rendered, flag it to a developer.
!Audit category cards showing Bot Accessibility, Schema Markup, Robots.txt, Sitemap, LLMs.txt all passing, and Readability at 51%
How to read the results
The audit flags issues by category. Don't try to fix everything at once.
!Pages Analysis table showing page URLs with Has Schema, Crawlable, and Bot Accessible status columns
Fix in this order:
- Bot accessibility and crawlability — if AI can't reach the site, nothing else matters.
- Schema markup, once access is confirmed.
- Structure and readability. These usually involve dev or content work, so they take longer.
- LLMs.txt, last and optional. Quick to add, but no proven effect on citations.
Why this comes before topics and prompts
AEO Copilot runs on a sequence: access, then understanding, then authority, then optimization.
Topics and prompts tell you how AI talks about a brand. But if the technical layer is broken, those results aren't reliable. The model might be working from incomplete or misread content. You'd be optimizing based on a distorted picture.
Run the audit first. Fix what it flags. Then move to tracking.
Frequently asked questions
How long does the technical audit take to run?
30 to 60 seconds for the scan. You can also import a list of URLs if the site is too large to crawl in one pass.
What categories does the technical audit check?
Structure, crawlability, bot accessibility, schema markup, llms.txt, authority signals, and readability.
What order should I fix technical audit issues in?
Bot accessibility and crawlability first: if AI can't reach the site, nothing else matters. Then schema markup, then structure and readability, which usually need dev or content work. llms.txt comes last: it is optional and has no proven effect on citations.
Why does the audit check for AI-specific bots like GPTBot and ClaudeBot?
Some sites block them by accident, often through robots.txt rules written for SEO purposes years ago that were never revisited for AI crawlers.
Why run the technical audit before setting up topics and prompts?
If the technical layer is broken, topic and prompt results aren't reliable — the model may be working from incomplete or misread content, so you'd be optimizing based on a distorted picture.
Not logged in yet? Run a free audit on any URL to see what the scan covers before signing up. For a full breakdown of the methodology in an agency context, see How to run an AEO audit for a client and the core guide to AI visibility tools.