Technical AI Audit
The Technical Side of AI Search
The right crawlers, allowed
Search crawlers, live lookups and training crawlers do different jobs. The audit separates them, so you can see whether a blanket “block AI” rule has also taken you out of AI search.
Every page that matters
Assistants answer questions about rooms, restaurants and venues from the pages about them. We read those pages, not only the homepage, and check each one carries the schema that describes it.
Content that arrives without scripts
If the text arrives only after scripts run, most AI crawlers read an empty page. We count the words in the HTML as it is served, which is what they see.
Signs the site is current
Assistants prefer sources that are clearly current. We look at your news dates and sitemap to see when the site last changed.
Common Questions
It reads your homepage, robots.txt, sitemap and llms.txt, then finds your rooms, dining, meetings and events, spa, what’s on, offers, news, FAQ and contact pages and reads those too. Each page is checked for crawler access, text that arrives without JavaScript, the schema that describes it, and signs the site is kept up to date.
No. It checks the part you control directly, your own website. Whether an assistant recommends you also depends on what other sites say about you, such as reviews, press coverage, OTA listings and your Google Business Profile. A site that passes is ready to be recommended. It is not a guarantee that it will be.
No. GPTBot, ClaudeBot and Google-Extended collect text to train models. The crawlers that power AI search answers are different ones, such as OAI-SearchBot, Claude-SearchBot and PerplexityBot. You can block training and still appear in answers, as long as the search crawlers stay allowed.
Most AI crawlers read the HTML your server sends and do not run JavaScript. If your room descriptions, rates or schema are added by scripts after the page loads, those crawlers never see them. The audit reads pages the same way, so what it finds is what they find.
It is a proposed plain-text file that summarises your site for language models. No major assistant has confirmed that it reads one yet, so it carries little weight in the score. It takes very little time to write, which is why we still suggest it.
Usually because the site turned us away. Bot protection and firewalls often challenge automated visitors, and AI crawlers are automated visitors too. If we cannot read your homepage, it is worth checking your CDN or firewall settings for rules that block bots.
