AuthMind Joins IBM to Build the Future of Agentic AI Identity Security

I'm incredibly excited to share that AuthMind is joining IBM! Together, IBM and AuthMind will build the future of Agentic AI Identity Security, bringing AuthMind's deep identity observability and intelligence to the HashiCorp Vault portfolio and IBM's approach to securing AI agents from identity to action.

IBM is building a continuous security model for agents: register them, delegate authority, authorize access, audit actions, and revoke risky behavior. HashiCorp Vault provides the foundation for this with identity-based controls, secrets management, and short-lived credentials. AuthMind complements this by showing how every identity authenticates to Vault, retrieves secrets & credentials, and uses them across systems and applications.

IBM is sharing its view of the news today as well - IBM advances agent identity security with new Vault capabilities. Here I want to share ours: the vision that started AuthMind, why AI agents made it urgent, and where we're headed together.

Identity Was Always the Front Line

When we founded AuthMind in 2020, our vision was that identity would move to the forefront of security. Most identity tools provision identities but stop watching once access is granted, and that's exactly where the risks live. 

We built AuthMind around observability for identities, detecting and protecting what they actually do vs. what policies intend. That covers every AI agent, non-human identity (NHI), and human identity, known or unknown, along with the secrets they use or misuse. With HashiCorp Vault, that means tracking each identity from the moment it authenticates and retrieves a secret to how it uses that secret across systems and applications, so security teams can spot risks, threats, misuse, and unusual activity as they happen.

Agents Made Identity the New Frontier

The rise of Agentic AI proved our vision faster than anyone expected. Agents are being deployed faster than organizations can govern them, often operating in the shadows and with no human involved in each action. Each known AI Agent is a legitimate identity with authorized credentials, yet its behavior can shift as it picks up new tools, delegates tasks, or reaches for new resources. That's why agent identity security can't be a one-time decision, made when credentials are issued. It has to continuously connect what an agent is authorized to do with how it acts once deployed.

Bringing Identity Observability to IBM's Agent Security Vision

IBM’s approach reflects exactly what we’ve seen in the field: securing agents has to span their entire lifecycle, from identity to action, and not stop at the moment access is granted. AuthMind supports that vision by discovering shadow agents and identities operating outside governance and distinguishing an authorized agent acting outside its scope from a compromised one, so teams can restrict or revoke access. As agents multiply, static permissions and manual processes won’t keep pace, which is why continuous identity observability belongs at the core of how IBM secures agents.

Special thanks to our partners and lead investors who supported us from day one Ballistic Ventures, Cheyenne Ventures, IBM Ventures, Blu Ventures,  Black Opal Ventures, K2 Access Funds, Jefferies Family Office, and F5.  

Most importantly, thank you to our incredible AuthMind team that made it all happen.

Read the Press Release - IBM advances agent identity security with new Vault capabilities