Scorecard for ecommerce and retail

In short

In the crawl of 4 Nov 2026, 94.5 percent of ecommerce sites with a readable robots.txt allowed the four main AI agents, and 83.4 percent of home pages showed readable text to a plain fetch. The sample is 256 domains, a complete first edition. Most rows come from Ora's E-commerce category.

The four numbers

  • 95%allow the four main AI agents
    190 of 201 sites with a readable robots.txt let GPTBot, ClaudeBot, PerplexityBot and Google-Extended in at the root
  • 83%show readable text on the home page
    176 of 211 home pages gave a plain fetch 150 words or more of text, not a script-only shell
  • 3.5%publish any act surface
    8 of 231 sites answered with an MCP manifest, an OpenAPI file, an agent card or agents.json at a standard path
  • 5.6%publish a payment or commerce file
    13 of 233 sites answered with a UCP profile at /.well-known/ucp.json

Against the whole crawl

Ecommerce and retail against all crawled domains
MeasureThis sectorAll crawled domains
Allow main agents94.5% (190 of 201)84.4% (7,282 of 8,624)
Readable home page83.4% (176 of 211)84.9% (7,457 of 8,783)
Any act surface3.5% (8 of 231)2.1% (216 of 10,385)
Any payment surface5.6% (13 of 233)0.6% (62 of 10,439)

The share allowing the main agents is 10.1 points above the crawl as a whole, and the share with readable home page text is 1.5 points below it.

How this sector was sampled

256 domains fall in this sector. The sector figures use Ora leaderboard categories as stand-ins, not a definition of the sector: E-commerce (153 rows), Commerce Tools (31 rows), plus 72 domains placed by a keyword on the domain name. 238 answered a plain fetch. The percentages use the denominators shown in each row, because a site with no readable robots.txt, or a home page that did not return 200, cannot be counted in that measure.

The methodology lists the category and keyword rules. Sector assignment by keyword is rough by design, and a site can sit in the wrong sector; the leaderboard category is the stronger signal. For context, Ora's own leaderboard averaged 52.7 out of 100 over 153 E-commerce entries on 5 Nov 2026 (vendor-published).

What to do with it

  • Ask your web team for the same four checks on your own site: robots.txt rules for the four agents, home page text without scripts, any act file, any payment file.
  • Take the ecommerce guide to the meeting where those decisions are made.
  • Run the self-check for ecommerce to see which of your own answers are open.

Sources

  1. Scorecard methodology: the crawl of 4 Nov 2026 (amireadyforagents.com/scorecards/methodology), read 4 Nov 2026.
  2. Sector scorecards dataset, edition 4 Nov 2026 (amireadyforagents.com/scorecards), read 4 Nov 2026.

Last reviewed