What is agent readiness?

Agent readiness measures how well a website can be read, cited, and interacted with by AI agents such as GPT-4o, Claude, and Perplexity. It is distinct from traditional SEO and requires different signals: structured data, semantic HTML, machine-readable content negotiation, OAuth discovery, and MCP compatibility.

What does the COVEN scanner measure?

The COVEN scanner measures five scored dimensions: Structure (how well an agent can parse your content), Citability (how likely an AI system is to cite your content), Discoverability (robots.txt, sitemaps, link headers), Agent Interface (MCP, OAuth discovery, API catalog, markdown negotiation), and Transactability (machine-readable commerce surfaces such as agent policy, structured offers, x402, MCP payment-capable tools). Agent Correlation is also measured but reported as a research signal that contributes 0% to the public score. Each scored dimension is weighted to produce a score from 0 to 100.

How is COVEN different from traditional SEO tools?

Traditional SEO tools measure visibility to search engine crawlers. COVEN measures whether AI agents can understand, cite, and interact with your site — which requires different signals and a different analytical framework. An SEO-optimised site can still score very poorly on agent readiness.

Yes. The single-page scan is free with no account required. You receive a score across all five dimensions and a list of specific issues to fix. Agent Analytics — ongoing monitoring and agent traffic detection — is also free to use during our three-month beta at aa.covenai.io.

Agent readiness · five dimensions · free scan

Is your site ready for AI agents?

AI agents are already visiting your website. Most sites are almost entirely unprepared. We built the tools to measure the gap.

Measurement, not middleware. Open rubric, paid tooling. Free during beta. Pricing earns its way in.

Open Agent Analytics → Run a Free Scan

How it works · free during beta, continuous monitoring, intelligence briefs.

83/100

Good

Dimension scores

Structure

Citability

Discoverability

100

Agent Interface

Transactability

Our flagship product

Agent Analytics — built for the agentic webBeta

Track how AI agents discover, cite, and transact with your site. Agent Analytics measures your presence across the five scored dimensions — Structure, Citability, Discoverability, Agent Interface, and Transactability — and surfaces the specific signals agents act on. Agent Correlation is reported alongside as a research signal. The free scan is a snapshot. This is the dashboard.

Open Agent Analytics → Free during beta · no credit card

Agent traffic share

See which AI agents visit your site, what they read, and how their behaviour compares to human traffic.

Citability across five dimensions

A continuously updated score with per-dimension breakdowns and the specific fixes that move each one.

Intelligence Briefs

AI-generated analysis of the gaps, wins, and changes agents are actually seeing on your site, week by week.

Historical trend analysis

Watch readiness and agent correlation move over time as you ship fixes — with alerts on meaningful changes.

Why this matters

AI agents are visiting your site. Most businesses have no idea.

When a user asks Claude, ChatGPT, or Perplexity to research a product or complete a task, the agent visits your site autonomously. It needs to understand your content, retrieve it cleanly, and discover what your site can do. Most websites give it almost nothing to work with.

Cloudflare scanned 200,000 of the most-visited domains in April 2026. Only 4% have Content Signal directives for AI agents.
BrightEdge data puts AI agent traffic at 15% of all web requests — and rising. Most analytics platforms do not track it.
The infrastructure for agent commerce (payments, auth, transactable APIs) is being laid right now. The window for getting ahead is closing.

From search visibility to machine legibility to transaction readiness — the funnel is changing.

The methodology

Five dimensions. One clear score.

Every scan measures the same five things. Each dimension is weighted by how much it affects an agent's ability to work with your site. The methodology is published openly.

Structure · 25%

Can an agent parse your content?

JSON-LD structured data, semantic HTML, heading hierarchy, and schema markup. Without these, agents have to guess what your pages mean.

Citability · 25%

Will an AI system cite you?

Publication dates, authorship signals, canonical URLs, and outbound citations. AI systems surface content they can confidently attribute.

Discoverability · 15%

Can agents find their way around?

Agents navigate by protocol. robots.txt with Content Signal directives, sitemaps, and RFC 8288 link headers tell agents what they are allowed to do and where to go.

Agent Interface · 20%

Can an agent interact with your site?

MCP server card, OAuth discovery (RFC 8414), API catalog (RFC 9727), and markdown content negotiation. This is the dimension that has no analogue in traditional SEO — and where most sites have nothing at all.

Transactability · 15%

Can agents buy from your site?

Whether the site exposes the machine-readable commerce surfaces purchasing agents look for — agent policy, structured offers, x402 payment shape, MCP payment-capable tools, DNS records announcing payment endpoints. We measure these signals; we don't transact for you. Open rubric, paid tooling.

Agent Correlation · research signal · 0% to public score. Observed agent traffic, dwell time, and return visits across the AI agent systems we track. Reported alongside the score for transparency, but contributes 0% to the public number. Treating it as scored would reward synthetic agent traffic over genuine agent-readiness.

Ongoing monitoring

Agent Analytics

See which AI agents are visiting your site, which pages they read, and how their behaviour compares to human traffic. Monitor your agent-readiness score over time as you make improvements.

JavaScript tag and server-side SDK for agent detection.
REST API for readiness scores, agent signals, and traffic data.
MCP-compatible interfaces for agent discovery.
Free to use during our three-month beta — open now.

Open Agent Analytics

Developers

Built for technical teams.

Instrument your stack, connect your data, and build toward agent-native infrastructure with COVEN's API and developer tooling.

REST API for readiness scores and agent signals.
MCP-compatible interfaces for agent tool discovery.
Open methodology and scoring rubric under Apache 2.0.
Email [email protected] to discuss the spec.

Visit developer docs

Join the waitlist.

Agent Analytics is free to use during our beta period. Paid subscription plans — with higher limits, team access, and priority support — open after beta in July 2026. Leave your email and we will let you know.

You are on the list. We will be in touch.

No spam. We will only email you when paid tiers open. Privacy Policy

See where your site stands.

Free scan. No account required. Five dimensions. One score. A list of exactly what to fix.

Run a Free Scan