Retail giants

Amazon agent readiness report

amazon.com · re-scanned every Monday

25
F

This site is essentially invisible to AI agents.

#12 of 15 in the library

Scanned 2026-07-08

Score history

min 25 max 25

Amazon is the default reference point for online retail, and as shopping agents start comparing prices and placing orders on a person's behalf, whether they can actually read an Amazon page cleanly matters as much as whether Amazon allows them in at all. Explicit crawler permissions are a start, not a finish.

The scan found amazon.com naming GPTBot, ClaudeBot, PerplexityBot, and Google-Extended directly in robots.txt, and the homepage does load for a scanner user agent over HTTPS. Past that, the page offers very little structure: no JSON-LD anywhere, no sitemap reachable at the standard location or referenced in robots.txt, no agents.md, llms.txt, or llms-full.txt, and no markdown response when one is requested. Meta hygiene is incomplete, missing Open Graph tags, and the text-to-markup ratio is far below threshold, with the scan noting that key content only appears after JavaScript runs.

Naming AI crawlers by name in robots.txt signals awareness of the problem. What is missing is everything that would let an agent act on that permission: a summary file, structured data, a non-JavaScript path to the actual content. Right now the invitation is not backed up by anything to read.

What Amazon gets right

  • robots.txt explicitly names and permits GPTBot, ClaudeBot, PerplexityBot, and Google-Extended
  • Clean HTTPS with forced HTTP-to-HTTPS redirect
  • Homepage accessible to a scanner user agent

What is missing

  • No agents.md, llms.txt, or llms-full.txt: assistants have nothing to read about how amazon.com wants to be used or summarized
  • No structured data: no JSON-LD found on the homepage, so agents cannot reliably extract what is being offered
  • No sitemap found at the standard location or referenced in robots.txt: crawlers have no map of the site, so new pages surface late or never
  • No content negotiation: requesting markdown still returns raw HTML
  • Meta hygiene incomplete: Open Graph tags missing, so descriptions elsewhere are improvised from whatever is available
  • Key content loads only after JavaScript runs: agents that do not execute scripts see a largely blank page
  • No MCP server card and no .well-known/verified.md: no way to interact with the site as a service, and no way to confirm it is authentic
  • No x402 payment signals detected

3 of 14 checks passing

See the full check-by-check report

Market context

AI traffic to US retail sites grew 393% year over year in Q1 2026, and AI-sourced shoppers converted 42% better than non-AI traffic. Source: Adobe Analytics, Apr 2026.

Your site can outscore Amazon in 5 days.

95 or higher, guaranteed, or your money back.

Scan for free See how we work Get a tailored proposal