Guidelines  ·  2026-06-08

OWASP Launches Enterprise Adoption Maturity Model for Agentic AI — Six-Level Deployment Axis vs Four-Level Governance Axis

GuidelinesMedium impactGlobal
At the OWASP GenAI Security Summit during Infosecurity Europe (June 4, 2026), the OWASP GenAI Security Project introduced the Enterprise Adoption Maturity Model for agentic AI, drawn from its June 3 paper 'State of Agentic AI Security and Governance v2.01'. The model maps six agent deployment levels (AT0 shadow AI through AT5 custom in-house agents) against four governance maturity levels (ad hoc through continuous governance-as-code), producing a red/yellow/green matrix that flags where governance fails to match deployment. OWASP also announced the formation of an Agentic Research Council to coordinate ongoing security research for agentic systems.
Most organisations deploying agents are operating in the 'red cells' — shipping AT3–AT5 autonomous agents while still running AT1-grade governance designed for AI copilots. The maturity model provides shared language for threat modelling, procurement requirements, and audit criteria that consulting teams and CISOs can apply immediately to justify specific control investments or autonomy reductions.
Plot your organisation's deployed agents on the OWASP deployment-governance matrix; any agent in a red cell must either gain named-owner accountability, AI-SBOM, real-time logging, and autonomy limits (to reach Level 2), or have its tool scope and permissions reduced until existing controls suffice.
Sources
OWASP GenAI Security Project — State of Agentic AI Security and GovernanceInfosecurity Magazine — OWASP Introduces Agentic AI Security Maturity Framework
See this in the live feed Explore related AI security and governance findings — updated every morning.
Open the feed →