What Is Agent Readiness? The Technical Standards That Make Websites AI-Operable
Ryan KingsFounder & CTOPublished June 2026 · 8 min read
Corrected 2026-11-04: earlier versions quoted Cloudflare, Lighthouse and FORKOFF scores that came from our own approximation of those tools, not from the tools themselves.
What Is Agent Readiness? The Technical Standards That Make Websites AI-Operable
Agent readiness is the set of technical standards that let autonomous AI agents discover, understand, and operate a website. It covers protocols, manifests, middleware, and discovery files across three layers: discoverability, content accessibility, and operability.
Unlike AI content readiness (whether answer engines cite your pages) or traditional SEO (human search rankings), agent readiness determines whether software agents can interact with your site at all. A site that publishes the files agents look for can be found through standardised manifests, queried for clean content, and operated through defined protocols.
Why does agent readiness matter in 2026?
Agent readiness matters because autonomous AI agents now evaluate specific technical signals before interacting with a website. If those signals are missing, agents skip the site or fall back to unreliable HTML scraping.
Two public checks cover parts of this. Cloudflare's isitagentready.com groups its checks into five categories: Discoverability, Content Accessibility, Bot Access Control, Protocol Discovery, and Commerce. A site-type control can limit a run to all checks, a content site, or an API. Google's Lighthouse Agentic Browsing category is experimental and does not produce a 0–100 score. It reports a pass ratio plus audits for the agent accessibility tree, WebMCP tool registration and schema validity, CLS, llms.txt, and an ARD manifest checked against ARD v0.91.
What are the three layers of agent readiness?
Agent readiness breaks into three layers, each building on the last: discoverability (can agents find you), content accessibility (can agents read you cleanly), and operability (can agents take actions on your site).
Layer 1: Discoverability. Agents begin with discovery files, not HTML crawls. robots.txt applies to every crawler through User-agent: * unless a bot has its own group. Explicit groups are only needed to treat a bot differently. The other files include a sitemap.xml referenced from robots.txt, an llms.txt file whose only required line is an H1 (a summary blockquote and link sections are recommended; Google Search does not use it), HTTP Link headers on the homepage, and auto-discovery link tags in the page head.
Layer 2: Content accessibility. When agents fetch a page, they prefer structured content over rendered HTML. Markdown content negotiation lets agents request clean text via the Accept header. Server-side rendering ensures content is visible without JavaScript execution. Schema.org JSON-LD blocks (Organization, WebSite, Article, FAQPage) help agents disambiguate entities. A correct H1-to-H6 heading hierarchy with stable IDs enables deep-linking to specific claims.
Layer 3: Operability. This is the frontier layer. It covers protocols that let agents act on your site: agents.json at /.well-known/agents.json, MCP Server Cards listed in /.well-known/ai-catalog.json (the card itself lives at the server's URL plus /server-card and does not list tools; see SEP-2127), A2A Agent Cards that declare supportedInterfaces, OAuth/OIDC discovery for authenticated workflows, and API Catalogs per RFC 9727.
What do the checks measure?
The checks are deterministic: each is a real HTTP probe, DNS lookup, or HTML parse, not a model-generated estimate. AEOForge's agent readiness tool sends requests as 5 different agent user-agents, parses robots.txt directives, resolves well-known endpoints, validates JSON-LD, and checks content negotiation responses. Every result is reproducible and re-checkable after fixes.
Other published checks, for cross-reference:
- Cloudflare isitagentready.com — five categories: Discoverability, Content Accessibility, Bot Access Control, Protocol Discovery, and Commerce. A site-type control can limit the run.
- Google Lighthouse Agentic Browsing — experimental, with no 0–100 score. Audits cover the agent accessibility tree, WebMCP tool registration and schema validity, CLS, llms.txt, and an ARD manifest checked against ARD v0.91.
What does a current check of our own site show?
The June 2026 write-up stated a before figure from an earlier version of these checks. That version is what this correction replaced. This page does not repeat it, and it does not quote a Cloudflare, Lighthouse, or FORKOFF score.
On 2026-11-04 the repaired check of aeoforged.com reported: Protocols found: 19 of 27 that apply to this site. Read score 100. 5/5 agent probes reached the homepage; discovery 4/4 (robots access, sitemap, machine-readable content, structured data).
The entire remediation took one working day using AEOForge's own tooling. The before and after scores are from the same deterministic probe suite, so the comparison is apples-to-apples.
Where should you start fixing agent readiness?
Start with discoverability signals because they gate everything else. If agents cannot find your manifests, operability protocols are irrelevant.
The recommended remediation order for 2026:
- robots.txt with AI bot rules and Content-Signal directive (10 minutes)
- sitemap.xml confirmed valid and referenced from robots.txt (10 minutes)
- llms.txt with an H1. A summary blockquote and link sections are recommended. Google Search does not use it. (1 hour)
- Auto-discovery link tags in your layout head (15 minutes)
- Markdown content negotiation via a single middleware file (1-2 hours)
- Schema.org JSON-LD for Organization and WebSite on the homepage (1 hour)
- agents.json + MCP Server Card generated from your live site and deployed to well-known endpoints (1-2 hours)
Steps 1 through 4 are file and header changes and can be done in under two hours. Full operability (steps 5 through 7) typically takes an additional half-day.
How do you measure progress?
Run an agent readiness check before and after each change. The probe returns which checks passed and which failed. Because the checks are HTTP-based and reproducible, you get an objective diff, not a subjective assessment.
Key takeaways
- Agent readiness is the set of technical checks that determine whether AI agents can discover, read, and operate a website.
- Cloudflare's checker uses five categories. Lighthouse's Agentic Browsing category is experimental and does not produce a 0–100 score.
- The three layers are discoverability (manifests and robots.txt), content accessibility (markdown negotiation and schema), and operability (MCP, A2A, agents.json).
- Every check is a deterministic HTTP probe or HTML parse, not an AI-generated estimate, so results are reproducible.
- On 2026-11-04 the repaired check of aeoforged.com found 19 of 27 applicable protocols, with a read score of 100.
- Start with discoverability: robots.txt, sitemap, and llms.txt. Those files can be published in under two hours.
- Agent readiness is distinct from content readiness and traditional SEO; a site can rank well for humans and still be invisible to agents.