Guide

What an AI agent readiness scanner measures.

An AI agent readiness scanner checks whether agents can reach, understand, navigate, and hand off from a website without a human reverse-engineering the page first. Siteline calls those four fundamentals SNAP: Signal, Navigate, Absorb, and Perform.

Signal

Signal asks whether agents can reach the site at all. HTTPS, server reachability, robots policy, and bot-protection behavior all matter. A site that blocks ordinary diagnostic crawlers may be invisible to useful agents even if it works in a browser.

Navigate

Navigate asks whether an agent can identify the organization, the important routes, and the task path. Clear page titles, structured navigation, sitemap coverage, and machine-readable discovery files all reduce ambiguity.

Absorb

Absorb asks whether the page content is readable as content, not just as layout. Server-rendered text, semantic headings, useful metadata, and content that survives agent-oriented transformations matter here.

Perform

Perform asks whether an agent can hand the human to a next step: contact, book, buy, subscribe, contribute, or escalate. The scanner is not trying to complete private transactions. It checks whether public handoff paths are interpretable.

Why the grade is capped

Siteline also checks active agentic enablement: files like llms.txt, agents.json, OpenAPI, changelogs, feeds, and security contacts. A site can be passively usable but still lack dedicated agent resources, so this second layer caps the maximum grade.