Compatibility check.

Your headline number is Agent Readiness (0–100): how well an AI agent can find, read and complete a task on your site. Technical Readiness (0–100) sits beside it — the passive checks alone, the same kind of evidence the open-source Scovant Core collects.

Both come from one R2 methodology evaluated at different evidence levels: Technical Readiness from the open passive checks; Estimated Agent Readiness when only passive evidence exists; Verified Agent Readiness once at least one agent journey has run on the site. Every result says which it is (readiness.agent.evidence_state: NOT_MEASURED, ESTIMATED or VERIFIED). A fourth state, ACCESS_REFUSED, marks a site whose every fetched page was a bot challenge: public pages then say "Automated access refused — rating not computed" and show no number or badge, while the site owner still sees the result.

Every rule in the public catalog (the catalog states which rules are scored and which are experimental — experimental findings are reported but never change the score) (Discoverability, Structured Data, Consistency, Actionability, Trust, UCP, Citability) are evaluated on every scan — the full, versioned rule catalog is public at /rules, including each rule's severity and maturity (required rules score; experimental rules are reported but not yet scored). Every check produces an outcome — PASS, WARN or FAIL when it could be measured, N/A when it does not apply, and "not measured" when its input was not available. A category's score is the weighted share of its measured checks that pass (a WARN counts half; a check's weight follows its severity, critical 40 down to info 2; a small fixed prior keeps one failure from driving a sparse category to 0), and the overall rating is a weighted blend of the categories that have a score. Missing evidence is never counted as a failure — it lowers coverage instead. The methodology is open — category weights, rule weights and outcome values are published on How scoring works — while the calibration layer (exact critical-cap mechanics, anti-gaming heuristics) stays proprietary: transparent score, proprietary engine.

Categories apply conditionally. UCP is scored only for commerce-like sites (commerce, booking, restaurant, SaaS); Citability only for content sites (blog, news, media, docs/documentation, portfolio, other). Where a category doesn't apply it is excluded and the remaining categories take its place — a blog is never judged on UCP, and a shop is never judged on prose citability.

Rules apply by capability. Product, offer and price rules apply only to a site with a product catalog; shipping and returns only to a site that sells physical goods; checkout rules only to a site that checks out. Scovant sets a capability when a scanned page shows it (a Product with an offer, an add-to-cart control, shipping details in the schema, a declared checkout) and otherwise uses the default for the site type — for example, a hotel or restaurant has no physical goods. When neither is known, those rules are marked not applicable rather than failed. On a SaaS site an add-to-cart or "buy now" control sells a licence or a subscription, so only Product data with an offer establishes a catalog there.

Agent access. A site that actively blocks AI agents (blocked AI crawler or bot-protection walls) cannot be agent ready while agents are locked out: its Agent Readiness is held below the compatible threshold and it earns no badge.

Commerce-tier sites (online stores, booking services, restaurants) also receive simulation-aware scoring: several AI agents attempt the purchase journey, and your readiness reflects how many of them complete it — when too few can, Agent Readiness drops and the badge is withheld, while agents completing the purchase helps it.

Badges: agent_verified at a Verified Agent Readiness rating of 80 or above, agent_compatible at 60 or above, estimated for a rating that rests on passive evidence only, none below 60 or without a number. Coverage: a result whose coverage is below 85% shows its number without a grade or badge (incomplete coverage). When the cause is the site itself, the scan says which: site refused our fetch when its firewall answered our scanner with an access refusal (for example HTTP 403 or 429) on every page, site returned errors when its pages answered with errors (404 or 5xx) — the rating then covers what we could measure; below 60% there is no number at all (not enough evidence). Grades: A at 90 and above, B at 80 (the agent-ready floor), C at 60 (the agent-compatible floor), D below — plus a confidence level (HIGH for a completed scan with domain-level probes, MEDIUM otherwise).

Your Agent Readiness rating and badge appear on your public showcase card and in the showcase API response (GET /api/public/showcase/{site_id}). For the full customer-facing explanation, see How scoring works.