Today I spent some time analyzing why ChatGPT, Claude and Perplexity keep recommending products like Supabase. It’s not only about having a good product. They optimized their entire web presence for AI discoverability: structured docs consistent terminology AI-readable examples crawlable open-source repos llms.txt and agent-oriented content Basically: LLMs can clearly understand what Supabase does, so they confidently recommend it. I think this becomes a huge distribution advantage in the next few years. That’s also why we built: LLMScan.dev A tool that audits how discoverable and understandable your company is for AI systems.

A few weeks ago I asked ChatGPT to describe a SaaS product I know well. It got the positioning half-right, mentioned a feature removed a year ago, and completely missed the main use case. The site looked good. Nice design, solid copy, decent SEO. But the information wasn’t structured clearly enough for AI systems to understand it properly. That’s becoming a real problem. When people ask ChatGPT, Claude, or Perplexity which tool to use, your website is no longer just being read by humans and Google. It’s being interpreted by systems trying to extract what you do, who it’s for, and whether it should recommend you. If that information is vague or buried, the output reflects it. That’s why I built LLM Scan. You paste a URL and get a plain-English report showing what AI systems struggle to understand about your product and how to improve it. Scan your site here: llmscan.dev Curious to hear what results you get and your feedback.

Presenting LLM Scan, a tool that allows you to improve your agents and llms discoverability. You get an audit and actionable feedback to improve your LLM readiness.