Agent readiness: what it is and how to measure it
Agent readiness is how well an online store can be reached, read and used by AI shopping agents that act for a shopper. It is not a ranking and not a promise of being recommended: it is a set of conditions you can check and fix.
What agent readiness means
An AI shopping agent visits your store for someone: it looks for a product, compares price, stock and terms, and may go as far as the cart. A store is agent-ready when the agent can get in, read the facts it needs in a form it can parse, and take the next step without a human stepping in. Each of those can be tested from outside, page by page.
Three layers
Our free test groups its 27 checks into 5 categories; they fall into three layers.
- Access
- Search and shopping agents are allowed in robots.txt and the security layer doesn’t turn them away; the product page is indexable and listed in the sitemap.
- Readability
- Product name, price, currency and stock are in the first HTML; structured data agrees with the page; shipping and return terms can be read by a machine.
- Action
- An agent can add to cart from the page and reach checkout without being forced to create an account.
Why it matters now
51% of US consumers used at least one AI tool to shop in the past month.
AI-referred visits to US retail sites grew +138% in May 2026 versus a year earlier.
Source: Adobe Analytics (Digital Commerce 360) ↗ ·
Readiness, AI visibility and SEO
| Agent readiness | AI visibility | Classic SEO | |
|---|---|---|---|
| Question it answers | Can an agent reach, read and use my store? | Do AI assistants mention or cite my brand? | Do my pages rank in search results? |
| How it is measured | Fixed checks on your public pages and a real browser run | Repeated questions to AI assistants | Rankings, crawl and index reports |
| What you change | Access rules, product data, policies, cart path | Content, sources and reputation | Content, links and technical SEO |
How to measure it
- Free test: 27 checks on one product page and the pages around it, a score from 0 to 100, and critical blockers that cap the score.
- Agent simulator: AI agents shop on your site in a real browser and stop at checkout; it shows where they get stuck.
- Agent Readiness Index: the same kind of checks across 836 stores, published with the method and its limits.
Run the free test on your store → How the free test works → · Agent simulator → · Agent Readiness Index 2027 →
What it isn’t
- A guarantee of being recommended: no one can promise what an AI assistant will suggest.
- A ranking of your store against others.
- A one-time label: stores and agents change, so readiness is measured again after each fix.
Questions
Is agent readiness the same as SEO for AI?
No. AI visibility asks whether assistants mention you; agent readiness asks whether an agent that arrives can read your store and act on it. You can be mentioned and still lose the shopper at checkout.
Do I need new protocols like UCP or MCP?
Not to start. Most gaps are in access rules, product data and the cart path; protocols help once those basics work, and our protocol map shows what is checked today.
What does a first check need?
Only your store address. The free test reads your public pages; it needs no sign-up and no access to your systems.