[REVISION HISTORY]

Okta AI agent security initiatives

Updated 2 times since CLSTR started tracking revisions of this situation.

What changed

2026-07-31 02:25 UTC → 2026-08-01 23:44 UTC · added ~~removed~~

In late July 2026 Okta ~~launched~~ introduced runtime governance tools that let enterprises register AI agents as identities, enforce real‑time policies and continuously certify their access, citing compliance with German regulations such as NIS2, GDPR and DORA. A week later the ~~company’s~~ Global CISO Insights report highlighted governance gaps in AI‑agent deployments, especially among German firms. To close those gaps Okta announced the acquisition of AI‑security startup Permiso. On 30 July 2026 Okta confirmed a ~~definitive~~ definitive, almost‑all‑cash agreement to buy Permiso for roughly $200 ~~million in an almost‑all‑cash deal,~~ million, with closing expected ~~to close~~ in Q3 FY 2027. ~~Permiso’s technology~~ Permiso, founded by former FireEye executives Paul Nguyen and Jason Martin, monitors human, machine and AI‑driven identities, flags anomalous behaviour and ~~includes~~ offers a sandbox called SandyClaw for testing AI‑agent skills before deployment. The ~~purchase~~ startup’s platform provides more than 2,500 research‑driven risk signals, behavioural analytics and an automated AI‑identity response capability, addressing a Cloud Security Alliance‑reported non‑human‑to‑human identity ratio of 144:1 and the fact that only about 21 % of organisations have formal AI‑agent governance programs. The acquisition expands Okta’s identity‑security fabric ~~to cover machine~~ into identity‑threat detection and response (ITDR), unifying real‑time threat detection with identity posture management and giving enterprises deeper visibility into non‑human identities ~~alongside traditional users.~~ across multi‑cloud environments.

Versions

Only revisions since CLSTR began indexing content versions appear here. Select a version to see what changed compared to the one before it.