Scoring
The 5 agent-readiness levels: from Invisible to Agent-Native
What each of the five AI-agent readiness levels means and which score range it maps to.
Why levels
A 0-100 score is easy to measure, but “73” on its own says little. So it folds into one of 5 levels — a clear label for the state you’re in.
The five levels
| Level | Score | Name | What it means |
|---|---|---|---|
| 1 | 0-19 | Invisible | AI barely sees the site: no basic signals, content isn’t machine-readable |
| 2 | 20-39 | Discoverable | Crawlers find the site but understand it poorly — the minimum is there (robots/sitemap) |
| 3 | 40-59 | Readable | Content is structured (Schema.org etc.), AI can extract it |
| 4 | 60-79 | Interactive | Agent interfaces exist (ai-agent.json, early MCP/OAuth) |
| 5 | 80-100 | Agent-Native | Fully ready: content + discovery + interactivity for agents |
How to move up
Each level is the next layer:
- → Discoverable: basic discovery — robots.txt, sitemap.
- → Readable: machine-readable content — Schema.org, llms.txt.
- → Interactive: agent interfaces — ai-agent.json, MCP.
- → Agent-Native: close the remaining gaps across all categories.
More in how to improve the score.
Find out your level
A free scan shows your score and level right away, with a per-category breakdown.