The largest names in identity and endpoint security moved within a single day to claim AI-agent identity as their territory, and several are acquiring their way in. The category a separate group of vendors entered this week is now drawing the industry’s biggest platforms, along with its first real wave of consolidation.

CrowdStrike introduced a Falcon Next-Gen Identity Security capability it calls continuous identity for AI agents, which authorizes an agent’s actions on an ongoing basis using the owner, the calling user, device posture, and real-time risk, rather than a single up-front grant. The capability builds on the company’s earlier acquisition of SGNL, a continuous-authorization vendor.

SailPoint said it plans to acquire Entro, adding non-human identity scanning, credential discovery, and machine-identity security to its Agentic Fabric platform, and said the deal would expand coverage to more than 1,000 types of non-human identity and agent. 1Password acquired Apono and launched a Credential Broker to govern access for human, machine, and AI identities, a move that carries the company beyond password management into just-in-time access governance.

The same day, Akamai unveiled an agentic security framework for its Bot and Agent Control products, linking agent identity, permissioning, traffic analysis, and edge enforcement. The framework incorporates Visa’s Trusted Agent Protocol and the Know Your Agent concepts being developed by Skyfire and Experian, with identity work from Auth0 and Ping. Saviynt expanded its Agent Access Gateway with intent-aware runtime authorization, which evaluates an agent’s actions in real time against identity, context, policy, and intent, and added verification features aimed at impersonation. Beyond Identity launched Ceros, a platform for unified identity, observability, and governance across agents and workflows.

The notable feature of the day is how much of it ran through acquisitions. Entro, Apono, and CrowdStrike’s earlier SGNL have all been folded into larger identity platforms, which suggests the fastest route to an agent-identity product right now is to buy one rather than build it. The category is consolidating at the same time it is forming.

Underneath the product packaging is a standards effort that has been running for months, as FIDO, the OpenID Foundation, and others work out how to authenticate an agent and describe what it is allowed to do. The commercial market has now caught up to that discussion and turned it into platforms, frameworks, and deals.

The shared premise across every announcement is the one the identity industry has settled on quickly: an AI agent should not inherit the permissions of the user, service account, or application that invokes it. It should be a governed actor with an owner, a policy, access boundaries, and an audit trail. A week ago that was a thesis a handful of vendors were testing.

Sources: CrowdStrike, SailPoint, 1Password

–
By the Control Plane Editorial Team