agent readiness score
Multiple signals, spread across at least two different axes: your site is starting to be designed to be understood and used by an AI agent, not just a human visitor.
Show this site's tier on your repo or page.
[](https://canagentsbuyfromyou.com/score/allbirds.com)
<a href="https://canagentsbuyfromyou.com/score/allbirds.com"><img src="https://canagentsbuyfromyou.com/score/allbirds.com/badge.svg" alt="agent readiness score: Agent Ready"></a>
https://canagentsbuyfromyou.com/score/allbirds.com/badge.svg
Readable and usable by an AI agent. Still to strengthen: Content accessibility.
Public snapshot · campaign from October 2, 2026
The 7 questions an AI agent asks on arriving at this site, and where it gets — or doesn't get — an answer.
We check that your site clearly signals its content and structure to the crawlers that visit it — like a floor plan handed out at the entrance.
Some sites unintentionally block AI crawlers, like an overzealous bouncer at the door. We check the door isn't shut by mistake.
We check that your page content is readable and usable by an AI — not just nice to look at for a human.
Beyond simple reading, some agents can connect to your site through dedicated channels — like a direct line instead of a regular visitor at the counter. We check whether those lines exist.
To act on your behalf (for example, to make a purchase), an agent needs to prove its identity securely — like a keycard rather than a password scribbled on a sticky note.
We simulate an agent trying to find a product, understand it, then buy it — how far does it get before it has to call in a human?
An agent gives up faster than a human when a page is slow or shifts under it while loading. We check the technical stability of your pages.
What a shopping agent finds when it queries this site through the commerce protocol.
Your store lets agents find products, prepare an order and track it: it is correctly usable by shopping agents. Monitor this setup over time to catch any regression.
Insufficient catalog completeness (30% over 12 sampled products): a shopping agent risks blocking or relying on incorrect information. Most frequent issues — prices shown in inconsistent currencies on 100% of products; descriptions that are missing, too short or not useful on 11% of products. Fix these fields as a priority before republishing the catalog for AI agents. Score capped at 30 (vs 99 on content quality alone): at least one presence criterion fails on more than 50% of sampled products — prices shown in inconsistent currencies. A shopping agent cannot transact reliably on this catalog regardless of content quality; these fields must be fixed as an absolute priority. Additional diagnostic: the description field is present but its content is insufficiently detailed or too promotional for a shopping agent to present it with confidence — enrich it with factual information (material, dimensions, use, care) rather than marketing hooks. Readability of your product pages for an agent that reads them (a channel separate from the catalog): 66% completeness over 3 pages read. Most frequently missing information: return policy, delivery times and costs, physical attributes (color, size, material, pattern). No effect on the score above.
A few sites in the “E-commerce — retailers & brands (D2C)” category: