Seven in ten business websites have never told AI crawlers anything at all

We audited 387 business websites before any optimization touched them. The question was simple: has this site made any decision - allow, block, anything - about the AI crawlers that now read the web?

69% have made no decision at all. Their robots.txt never mentions GPTBot, ClaudeBot, PerplexityBot or Google-Extended - and for 8.5% of sites that's because there is no robots.txt to mention them in. These sites are crawlable by default, invisible by design to no one, and governed by nothing.

The rest split two ways. 12.7% explicitly block GPTBot - some deliberately, many via a security plugin or a copied template the owner has never read, which means a meaningful slice of businesses are opted out of ChatGPT without knowing it. And roughly one in five have written some AI directive, the early adopters.

The supporting files tell the same story. 52% of audited sites have no llms.txt (and our panel skews toward owners already curious about AI visibility - the open web almost certainly sits higher). One in ten has no XML sitemap. The average robots.txt scores 48 out of 100 against AI-readiness checks.

The fix list is short and unglamorous: a robots.txt that names the AI crawlers and states your choice, an XML sitemap submitted where the AI engines actually look (Bing feeds several of them), an llms.txt as forward cover, and schema so machines read your pages the way you meant them. EZY's free tier generates and uploads all four on WordPress and Cloudflare in one click.