The agent-era
protocol stack.
Every protocol we score in the Agent Readiness Score, explained with copy-pasteable examples.
Task discoverability
Can an agent reach the actions a human visitor would? We check for contact, pricing, signup, login, demo, checkout, docs, and support paths — applicability depends on the site archetype.
FAQPage schema
If you have an FAQ section, mark it up with FAQPage + Question/Answer JSON-LD. Highest-leverage AEO win — directly fuels Google AI Overviews and ChatGPT search citations.
Schema.org coverage
JSON-LD structured data on each major page type — Organization, WebSite, Product, Service, FAQPage, BreadcrumbList, Article, etc. The fastest path to AI Overviews and rich answers.
AI bot directives in robots.txt
Explicit User-agent rules for AI crawlers (GPTBot, ClaudeBot, PerplexityBot, etc.) — whether allow or disallow. Demonstrates an intentional posture toward AI traffic.
Cloudflare Content Signals
A robots.txt directive (Content-Signal: search=yes, search-ai=yes, ai-train=no) declaring how your content may be used by search, AI search answers, and AI training.
Web Bot Auth signing directory
Publish /.well-known/http-message-signatures-directory with a JWKS-like list of public keys you'll accept on signed bot requests. Lets WAFs safely relax rules for verified bots.
noindex / nosnippet / max-snippet blockers
Meta robots tags (noindex, nosnippet, max-snippet:0) prevent search engines and agents from indexing/quoting your content. Sometimes intentional, sometimes accidental.
Business clarity
Can an AI agent answer 'what does this company do, for whom?' from the homepage alone? Vague taglines and inside-baseball jargon make you uncitable.
/.well-known/* capability discovery
RFC 8615 defines /.well-known/ as a reserved namespace for site-wide metadata. Agents probe a known set: oauth-authorization-server, openid-configuration, mcp.json, agents.json, api-catalog, etc.
Agent Skills index
Enumerable list of discrete skills your site exposes — lighter than MCP, heavier than a raw OpenAPI blob. Path: /.well-known/agent-skills/index.json.
MCP server card (.well-known/mcp.json)
Anthropic's Model Context Protocol manifest. Makes your product natively installable in Cursor, Claude Desktop, and ChatGPT — agents can discover and call your tools without a custom integration.
Agentic commerce protocols (ACP, UCP, MPP, x402)
Four overlapping standards that let AI agents pay and transact: Agentic Commerce Protocol, Universal Commerce Protocol, Merchant Payments Protocol, x402.
Image alt text
Every meaningful <img> needs descriptive alt text. Agents rely on alt to understand image content; missing alts hide product photos, screenshots, and diagrams from AI.
Markdown content negotiation
When an agent sends Accept: text/markdown, return the same page's content as Markdown — stripped of layout noise, CSS, and JavaScript.
Semantic HTML structure
Use <h1>-<h6>, <main>, <nav>, <article>, <section> instead of generic <div>s. Agents use heading hierarchy as the spine of their page understanding.
Server-rendered HTML
Critical content must be present in the initial HTML response. Most AI crawlers don't execute JavaScript; client-rendered pages return blank to them.
llms.txt
A Markdown file at /llms.txt that gives agents a curated, human-edited map of your site's most important content — distinct from a comprehensive sitemap.xml.
CTA clarity
Buttons should describe the outcome, not the mechanism. 'Start free trial' beats 'Click here'. 'Get a price' beats 'Submit'. Agents parse CTA text to decide what action they're taking on behalf of the user.
RFC 8288 Link headers
Send Link headers on HTML responses to point agents at related resources (sitemap, llms.txt, MCP card, OpenAPI) without requiring them to parse HTML first.
Sitemap declared in robots.txt
robots.txt should declare your sitemap URL via 'Sitemap: <absolute-url>'. Ensures agents that read robots first can find the sitemap without guessing.
XML sitemap
An XML sitemap enumerates the URLs you want crawled. Agents prefer to seed their crawl queue from a sitemap rather than guess at links.
robots.txt
Every crawlable site should publish a robots.txt at the origin root. AI agents read it to learn which paths are allowed and which user-agents are blocked.
Entity consistency
Company name, logo URL, and sameAs links to LinkedIn / X / GitHub / Wikipedia / Crunchbase should match exactly across schema, footer, and your social profiles.
Content freshness signals
Articles, blog posts, and news need datePublished + dateModified. Agents discount stale content; explicit freshness dates earn citations on time-sensitive queries.
Trust signals
Customer logos, testimonials, case studies, security badges (SOC 2, ISO 27001), press mentions. Agents weight these heavily when ranking competing answers.