IBM AI agents can now be managed with generally available identity and access controls, the company announced on October 27, 2026. IBM also introduced tools to discover and isolate agents across platforms, although those features remain in technical preview.
Key Takeaways
- Agent Identity and HashiCorp Vault 2.2 are generally available, as are watsonx Orchestrate’s evaluation capabilities.
- Orchestrate’s agent discovery and containment features are in technical preview.
- Secure Hub, which provides visibility across Vault environments, is in public preview.
How IBM AI agents get access controls
IBM announced the capabilities at its TechXchange conference as businesses deploy AI agents that can use tools and interact with applications across different systems.
When used with HashiCorp Vault, Agent Identity provides a record of an agent’s owner, lifecycle status and granted authority. Vault is designed to issue short-lived credentials and limit access to the permissions an agent needs. IBM says access can be denied and a session revoked in real time if an agent attempts an unauthorized action.
Vault 2.2 also gives agents their own secured accounts, automatically renews credentials and incorporates post-quantum encryption, IBM said.
| Item | Details |
|---|---|
| Agent Identity and Vault 2.2 | Generally available |
| Orchestrate evaluation | Generally available |
| Orchestrate discovery and containment | Technical preview |
| Secure Hub | Public preview |
| Integration Intelligence Hub | Technical preview |
Discovery and containment remain in preview
IBM’s watsonx Orchestrate is designed to discover agents built on third-party platforms including Amazon Bedrock, Microsoft Foundry and Google Gemini. The discovery capability, which also extends to models and Model Context Protocol servers, is in technical preview.
A separate containment feature, also in technical preview, runs agents in isolated sandboxes and gives teams a kill switch to shut them down. The generally available evaluation tools let teams test quality, accuracy and compliance before deployment and monitor agents in production. Cost tracking shows model and token use by agent.
Other announced integrations
IBM said Agent Identity and Vault integrate with NVIDIA OpenShell, combining identity and credential controls with a protected runtime environment.
IBM Integration Intelligence Hub, which is in technical preview, is designed to give agents context about existing systems and reusable tools. IBM also listed Terraform policy as generally available, while its Developer Portal and remote MCP server are in public preview.
Read More News
Disclaimer
This FinanceVelix article provides general information, not personalized financial, investment, legal, or tax advice. Market prices, interest rates and economic conditions can change; investments carry a risk of loss. Facts reflect the material available when published and may change. Check important information independently and consult a qualified professional when needed. Mentioning an asset, company, or service is not an endorsement.
Comments
Be respectful. No spam, impersonation or private credentials. Ordinary posts and comments appear immediately; suspicious links are held for review.
Please wait…