AI assistants are already shopping. Are they buying from you, or from your competitor?
Find what stops AI shoppers from ordering on your store. Free, in 20 seconds.
AI shopping agents could drive up to 20% of US e-commerce by 2030 (Morgan Stanley). If an agent can't find, understand and check out on your store, you're invisible to it. Deterministic, not a vibe: no model grades your store, every check is a parseable fact, so the same store gets the same score, every time. A number that never moves on its own can be compared, ranked and cited by you, by your rivals, by the Index, and by the AI assistants that read our reports.
🔬 We test our own score against what AI agents actually do, on 54 stores, with the bar written before the data. Read the result and its limits. See the study →
Fetching your homepage like an AI shopping agent…
What the scan checks
Six pillars, from the most blocking to the most subtle. Four are scored today — access, structured data, catalog, transaction rails — by a deterministic engine (pure parsing, 100% reproducible: two scans give the same score). The last two are in preview and don't move your number yet.
1 · Access & crawlability
robots.txt allows GPTBot / ClaudeBot / PerplexityBot, llms.txt present, sitemap — plus two tests nobody else runs: the anti-cloaking differential, and the verified-agent check: our crawler cryptographically signs its requests (Web Bot Auth, the identity layer under Visa's and Mastercard's agent programmes) and detects whether your firewall recognizes verified agents.
2 · Structured data
Schema.org Product / Offer in JSON-LD: price, currency, availability, brand, SKU/GTIN, reviews. A pretty page with no JSON-LD is mute to an agent.
3 · Catalog & attributes
Variants and attributes (size, color, material) as fields, not prose. Without fields, an agent can't filter.
4 · Transaction rails & agent access
The most "2026" pillar, and it counts toward your score today: the UCP agent-checkout endpoint (/api/ucp/mcp), your A2A agent card, the new ARD ai-catalog, machine entry points on the homepage, returns & shipping in offers, a machine-visible add-to-cart. Almost no store passes these yet — early movers pull ahead.
5 · Real-time data preview
Up-to-date, page↔feed-consistent price/stock. We already measure silent drift — the facts agents buy on changing between scans with no visible signal; scoring lands once the checks are attestable.
6 · Logistics & trust first slice live
Machine-readable returns & shipping are already checked and scored (inside pillar 4). Next: delivery-time promises and structured reviews.
Pro is a console, not a report
The free scan is the diagnosis. Pro is the cockpit you fix and defend with — your priorities, your rivals, your whole catalog, watched daily. This is the actual product, on a demo store:
We hold ourselves to the same bar: agentready.market scores 88/100 on our own audit — see the live report. A real scan, not a brochure.
One-time payment · no auto-renewal — nothing to cancel · 14-day refund, no questions asked.
Why now: agentic commerce is the next channel
"Bot SEO" — making your store readable and buyable by AI agents — is going from a nice-to-have to a primary acquisition channel. The forecasts diverge on the exact number, but they all point the same way: up and to the right.
$200–500B
US agent-driven e-commerce by 2030 — roughly 10–25% of online sales. (Morgan Stanley, Bain)
$15T+
B2B trade mediated by agents by 2028 — the biggest pool of all. (Gartner)
$1T US / $3–5T global
Agent-orchestrated consumer revenue by 2030. (McKinsey)
+42% conversion
AI-referred shoppers convert ~42% better than other traffic. (Adobe, Q1 2026)
The payment rail is being commoditized by the giants (Visa/Mastercard/Amex, ACP, UCP — 1M+ Shopify
merchants auto-enrolled). The value migrates to the optimization / data layer: making
your store agent-readable. Discovery (AEO/GEO — "does ChatGPT cite me") is already saturated; the
transaction layer (AOO — "can an agent actually buy") is where we play.
Highly uncertain projections — figures diverge by scope. Not investment advice.
How it works
The scanner opens the door. No industrial scraping: a targeted, polite fetch of public pages — the same handful an AI shopping agent would read — and a store owner can leave the public Index with one email.
You submit your URL
We fetch your public pages (home, one product page, robots/llms/sitemap, feed) — no purchase, no account.
We parse, no AI
Deterministic core: every check returns a reproducible result, weighted into the score.
Score + fixes
A score /100, your vertical benchmark, and the 5 fixes ranked by impact × ease.
You fix, you climb
Each applied fix is a before/after measurement. The score becomes empirical.
Guardrail: we simulate up to the checkout boundary, never a real purchase, no auth bypass, no personal data stored. Public pages, GDPR-friendly. And you don't have to take our word for who's knocking: every request carries an audit header and a cryptographic signature (Web Bot Auth) you can check against our published key — verify our crawler.
Ready to see your score?
Most stores sit below 60. Where's yours?
Scan my store