Okta's $200M Permiso Bet: AI Agents May Need a New Identity Trade

Okta Is Paying Up for an Emerging AI Identity Category

Okta agreed to acquire Permiso in a deal valued at just under $200 million. Okta is betting on what AI identity security could become, rather than buying a category with settled pricing or proven revenue benchmarks. The company said the transaction is structured as an almost all-cash deal, and it is expected to close in the third quarter of fiscal 2027. Until that happens, this remains a live strategic bet rather than a completed integration.

What Okta is buying

Permiso's work centers on monitoring AI agents and other machine identities and helping security teams spot suspicious activity in cloud environments after users or applications have been granted access. That fits Okta's broader push beyond login verification and toward continuous monitoring of what users, applications, and AI agents do after they gain access.

This is still a young category. Okta is not just buying technology; it is trying to shape the controls that could define agent security before standards harden. If enterprise budgets shift that way, the deal could look early and decisive. If demand stays fragmented, the premium becomes harder to defend.

Okta for AI Agents Tries to Solve a Different Identity Problem

Okta's core argument is straightforward: AI agents are challenging traditional identity security practices. Its response is Okta for AI Agents, which it describes as a platform to discover and register known and unknown AI agents, standardize agent access, and instantly revoke access to limit damage from rogue agent behavior.

Why the control loop matters

That three-step loop is different from human-only identity workflows. Agents can appear temporarily, move across systems, and keep long-lived permissions. Okta's approach starts with visibility, moves to policy, and then relies on a fast kill switch if something goes wrong.

Okta is also leaning on distribution. Its pitch includes extending coverage across 8,200+ integrations to reach more AI agent platforms and workflows. The strategic upside is clear only if that reach helps customers actually find agents, govern them, and tighten controls in one place.

Why Okta has a chance, and where it could stumble

Okta already sells beyond basic identity, including API Access Management, Machine-to-Machine Tokens, Identity Security Posture Management, and Identity Threat Protection. If agent security becomes a foothold for deeper consumption of those modules, the business case strengthens.

The main watchpoints are simple: - Whether Okta for AI Agents becomes a reason customers buy more than a brochure layer. - Whether discovery leads to actual access control, not just visibility. - Whether the revoke-access capability matters in real incidents enough to support expansion across Okta's broader suite.

What Would Make the Permiso Deal Work

The cleanest test is adoption. Okta needs customers to use the full control loop, not just the "where are my agents" narrative. If buyers rely on discover, register, standardize, and revoke as operating controls, the acquisition starts to justify itself.

If agent security remains mostly exploratory, or enterprises fold it into existing identity hygiene without paying a premium, the strategic logic weakens. The same goes for any monetization story that depends on deeper add-on consumption without proof that customers actually want that bundle.