This complaint shows up weekly in SEO communities, and it has a precise technical answer: Google rankings measure backlinks and relevance for a crawler that indexes everything. AI assistants are a different pipeline with different bottlenecks. Four of them matter:
1. Your robots.txt blocks the right bots by accident
Assistants fetch with their own user-agents (GPTBot, ClaudeBot, PerplexityBot). A wildcard block meant for scrapers often catches them too — and unlike Google, they respect it and move on silently. Check: curl -A "GPTBot" https://your-site.com/robots.txt and look for disallow rules touching your pages.
2. There is nothing machine-readable about you
Assistants summarize; they don't browse like humans. If your value proposition lives in JavaScript-rendered marketing copy, there's no cheap way to quote you. An /llms.txt file — plain Markdown stating who you are, what you sell, prices, buy links — gives them quotable facts.
3. You're not in their training or retrieval set yet
Assistants favor sources that are frequently cited, structured, and recent. Fresh, specific, factual pages (pricing tables, comparison pages, changelogs) get picked up far sooner than brand homepages. Publishing concrete content accelerates inclusion more than any meta tag.
4. Agents can read you but can't transact with you
The next wave is agentic buyers. Without an /.well-known/x402 manifest (or at minimum agent-legible checkout instructions), an assistant can recommend you but not complete a purchase — and recommendations without conversions get deprioritized over time.
Which of the four is blocking you?
Run the free audit — instant gradeOr track all surfaces weekly: AI Visibility Report, €19/mo