SEO-GEO · the 23SEOGEO team

Agent Readiness Auditor — make your site usable by AI agents | SEO-GEO

Traditional SEO asks 'can Googlebot read this?'. Agent Readiness asks 'can an AI agent call this?'. The former wants Schema/Lighthouse; the latter wants SearchAction/ReserveAction, autocomplete semantics, and an explicit AI bot section.

Google I/O 2026 (May) announced Search Agents (Information Agents shipping summer 2026), the Chrome WebMCP proposal, and live mini-app generation in Antigravity. Google's own framing: 'Search is moving from finding pages to completing tasks.' That means AI agents will preferentially call sites by the capabilities they expose — agent-ready sites win agent traffic first.

What can you check or produce?

  • 6-dimension 0-100 scoring: form semantics / callable actions / agent policy / structured price / action coverage / auth boundary
  • Batch audit up to 10 URLs, pure fetch — no AI cost
  • Parses robots.txt + llms.txt, detects 7 AI bot sections (GPTBot/ClaudeBot/PerplexityBot/Google-Extended/Bytespider, etc.)
  • Every issue ships with a one-click JSON-LD / robots / llms.txt fix template
  • Detects 5 high-value Schema actions: SearchAction / ReserveAction / OrderAction / BuyAction / SubscribeAction
  • All data persisted to sessionStorage — reload safe

How should you interpret results and choose the next action?

Google I/O 2026 (May) announced Search Agents (Information Agents shipping summer 2026), the Chrome WebMCP proposal, and live mini-app generation in Antigravity. Google's own framing: 'Search is moving from finding pages to completing tasks.' That means AI agents will preferentially call sites by the capabilities they expose — agent-ready sites win agent traffic first. Unlike a regular Schema audit that asks 'what is this page?', Agent Readiness asks 'what can this site do?' — can it search, reserve, order, subscribe? We pull evidence from raw HTML, JSON-LD potentialAction, and robots/llms.txt across 6 weighted dimensions. Pure edge-function implementation, User-Agent SEO-GEO-AgentReadinessBot/1.0, compliant crawling, free with your login account.

Frequently asked questions

How is this different from the Rich Results Auditor?

Rich Results Auditor checks whether a page's JSON-LD is complete (FAQPage/Article/Product, etc.). Agent Readiness checks whether the site's capabilities are exposed to agents (SearchAction/ReserveAction/Auth boundary/llms.txt). Complementary, not overlapping.

Scoring weights?

6 weighted dimensions: callable actions 22% + machine policy 20% + form semantics 18% + structured price 15% + action coverage 15% + auth boundary 10%. 85+ = A, 70-84 = B, 55-69 = C.

Is it paid?

Free for all logged-in users — pure fetch, no AI quota consumed.

WebMCP is still a proposal — why bother now?

Even before WebMCP standardizes, Google Information Agents already read SearchAction and ChatGPT can browse-and-act. The 6 signals this tool checks are all live today; WebMCP adoption just adds another bonus layer.

By · Last updated: