“Is my website ready for AI?” usually gets answered with a technical checklist: can GPTBot crawl you, do you have an llms.txt file, is your Schema.org markup in place. Those things matter, and we will get to them. But they are only half of the question.

The other half is older and costs you more: when a real person lands on your site with a question, can they get an answer? Because if they cannot, they leave and ask an AI instead, and you had better hope the AI knows your business well enough to answer for you.

Here are the seven checks we run when we audit a website for AI readiness, in the order they tend to cost you leads.

1. Can a visitor with a question get an answer?

This is the single biggest gap we see. A site describes its services, but there is no FAQ, no chat, no assistant. Every unanswered question becomes a bounce, an email you have to answer manually, or a visit to a competitor who explains things better.

The check: does the site have a live chat widget, an assistant, or at least a genuine FAQ? Any of these counts; having none of them is a high-severity finding.

2. How fast can someone find your contact details?

Contact information more than one click from the homepage might as well not exist for an impatient visitor. The check measures click depth: contact details on the homepage or one click away pass; buried or missing details fail.

3. Is pricing findable?

Hiding pricing is sometimes a deliberate sales strategy. More often it is an accident of site structure, and it filters out exactly the visitors who were ready to buy. The check: can pricing information be found at all, and how deep?

4. Speed, measured on a phone

Mobile performance is where slow sites lose people before a single word is read. We measure the Lighthouse mobile performance score and Largest Contentful Paint via Google PageSpeed Insights. A performance score below 60, or content that takes more than four seconds to paint, is a finding.

5. The SEO basics still matter

Title tag, meta description, a single H1, canonical URL, Open Graph tags, robots.txt, sitemap.xml. None of this is new advice, yet many sites are missing at least one of them. These basics feed classic search engines and AI search alike.

6. Can AI search engines read and cite you?

Now the technical half of AI readiness. Three things decide whether AI search products can use your site as a source:

  • AI crawler access. If your robots.txt blocks GPTBot, ClaudeBot, PerplexityBot or Google-Extended, the corresponding AI products cannot read your site, so they cannot cite it or recommend you. Sometimes that is a deliberate choice; for most businesses it is an accident inherited from a template.
  • llms.txt. An emerging convention: a plain-text file that tells AI systems what your site is about and where the important pages are. Few sites have one yet, which is exactly why having one is cheap differentiation.
  • Structured data. Schema.org markup (JSON-LD) tells machines unambiguously what your pages mean: your products, prices, opening hours, organization. Both classic and AI search rely on it. For a deeper look at how AI models actually consume your pages once they can reach them, see How AI Assistants Read Your Website (and What They Miss).

7. Cookie consent

Not glamorous, but if you serve EU visitors and set non-essential cookies, a consent banner is a legal requirement, and its absence is a trust signal savvy visitors notice.

Where AskMyWebsite fits

If these checks show that visitors have no way to get answers on your site, that is the problem AskMyWebsite exists to solve: an AI concierge that answers your visitors’ questions directly from your website’s existing content and shows them where to click. You do not rewrite anything; the concierge reads what is already there.

Work through the seven checks first. Whatever they show, you will know exactly where your website stands - with visitors and with the AI tools they are starting to trust.