State of Agent-Readiness 2026
How - enterprises measure up on the capabilities autonomous agents will actually need.
How to read a zero. Every figure on this page comes from an external scan of the public surface plus opt-in code analysis. A 0 or L0 means we found no public evidence of the capability - absence of evidence, not evidence of absence. 41 of the framework's 258 checks cannot be evidenced this way at all, and two dimensions (D16 human-agent collaboration, D18 threat defence) have almost no external reach.
Why is my organisation in this report?
Assessments in this benchmark use only publicly available information - public HTTP endpoints and public GitHub repositories. No authentication, no self-report, no private data. If you represent an assessed organisation and want your entry corrected or removed, contact us and we will suppress it from the public corpus.
See where you stand
Run a free assessment against 500 benchmarked organizations. 211 automated checks.
