IBM has announced new tools to give AI agents verified identities and tighter access controls, aimed at companies that are deploying agents faster than they can track them. Unveiled at TechXchange 2026 in Atlanta, the AI agent identity features extend HashiCorp Vault to agents, while IBM watsonx Orchestrate gains tools to discover, contain and evaluate agents built on other platforms.
IBM frames the problem with Gartner figures cited in its release: by 2028, the average Fortune 500 enterprise will use over 150,000 AI agents, yet only 13% of organisations believe they have the right AI agent governance in place.
Why AI agent identity matters
According to IBM, persistent credentials, unclear ownership and disconnected controls make agents hard to govern as they call tools, talk to other agents and act across teams and clouds. IBM describes agents as a growing class of non-human identity, alongside applications and services.
Its answer has two parts. Agent Identity, used with HashiCorp Vault, creates an authoritative record for each agent: who owns it, its lifecycle status and the authority it has been granted. Vault then uses that context to control the credentials the agent uses to reach databases, applications and cloud services.
Short-lived credentials and real-time revocation
IBM says an agent with a long-lived, overprivileged credential can put systems and data at risk if it is compromised or strays from its task. Vault will issue secured, short-lived credentials that limit each agent to what it needs. If an agent tries to retrieve information or act outside its approved permissions, IBM says access can be denied and the session revoked in real time.
Other announcements include:
- Vault 2.2: gives agents their own secured accounts, automatically renews their credentials and adds post-quantum encryption.
- Secure Hub (public preview): central visibility across a company’s Vault environments.
- NVIDIA OpenShell: Agent Identity and Vault integrate with NVIDIA’s secure runtime, pairing a protected environment with a verified identity and short-lived credentials for each agent.
IBM says Vault already secures non-human identities for organisations including ABN AMRO, Adobe, BlackRock, Canva, Expedia, London Stock Exchange, Starbucks and Target.
Managing agents across Bedrock, Foundry and Gemini
The second half of the announcement focuses on visibility. IBM says watsonx Orchestrate works across the AI ecosystem rather than inside a single vendor’s platform. Its new capabilities are:
- Discover (tech preview): connects to platforms such as Amazon Bedrock, Microsoft Foundry and Google Gemini to find agents built in third-party environments, as well as models and MCP servers.
- Contain (tech preview): runs agents in isolated sandboxes so a rogue agent cannot interfere with others, and adds a kill switch to shut down any agent.
- Evaluate: tests the quality, accuracy and compliance of agents, models and MCP servers before deployment, monitors them in production and tracks each agent’s model and token use to show what drives cost.
Orchestrate also works with Vault’s credential controls and Agent Identity, so organisations can authorise and audit agent actions and revoke trust when risk changes.
What is available now
According to IBM, Vault 2.2, Agent Identity and the Orchestrate evaluate capabilities are generally available. Secure Hub is in public preview, while the Orchestrate discover and contain capabilities, along with IBM Integration Intelligence Hub, are in tech preview. IBM also announced updates to IBM Concert, HashiCorp Terraform and a new IBM Z Atlas tool for mainframe change management.
“For agents to take on real work across the systems a company depends on, it requires an AI operating model,” said Dinesh Nirmal, Senior Vice President, IBM Software.
As a writer, I feel the kill switch and the 13% governance figure say a lot together: companies want agents doing real work, but most still lack a reliable way to see and stop them.
FAQ
What did IBM announce for AI agent identity?
IBM announced Agent Identity, which works with HashiCorp Vault to record each agent’s owner, status and authority, and to issue short-lived, least-privilege credentials that can be revoked in real time.
Which platforms can watsonx Orchestrate discover agents on?
IBM says Orchestrate’s Discover capability, in tech preview, connects to platforms such as Amazon Bedrock, Microsoft Foundry and Google Gemini to find agents built there.
Are these features available now?
Vault 2.2, Agent Identity and Orchestrate’s evaluate tools are generally available. Secure Hub is in public preview, and Orchestrate’s discover and contain tools are in tech preview.