AI & agents
IBM introduces identity and access controls for AI agents
- Published
- Source
- IBM Newsroom
Summary
At TechXchange 2026, IBM announced capabilities that give enterprise AI agents verified identities, least-privilege credentials, and traceable actions. The centerpiece, HashiCorp Vault 2.2, assigns each agent its own secured account, automatically renews credentials, and can revoke access in real time when an agent steps outside approved permissions. A 'Discover' tech preview in watsonx Orchestrate can even find agents built on third-party platforms such as Amazon Bedrock, Microsoft Foundry, and Google Gemini and manage them in one place.
Why it matters for our work
As agent counts grow, organizations that cannot answer who built an agent and what it may do lose the ability to govern it, and practical tooling for that problem is now arriving. With Gartner projecting an average of 150,000 AI agents per Fortune 500 company by 2028, governance infrastructure is becoming a precondition for how work gets done.
Translated from the Korean original. Summaries may be translated and edited. Commentary reflects our perspective; forecasts remain the source’s views.