New Identity and Security Solutions for AI Agents

IBM introduced new capabilities for the oversight and security of enterprise AI agents at the TechXchange 2026 event held in Atlanta.

IBM has announced new security and auditing features to ensure that AI agents are managed with clear authority and accountability. The initiative, unveiled at the TechXchange 2026 event in Atlanta, focuses on agent discovery, verified identities, access controls, and performance evaluation.

In this context, HashiCorp Vault integration is being extended to AI agents to provide secure, short-lived credentials, and an 'Agent ID' is being established to track ownership and lifecycle status. IBM watsonx Orchestrate has gained features to discover agents on third-party platforms such as Amazon Bedrock and Google Gemini, isolate rogue agents in sandboxes, and evaluate performance.

The new toolset also offers integration with NVIDIA OpenShell for secure runtime, while including updates to IBM Concert and IBM Z Atlas for infrastructure security and change management. These solutions aim to bridge the enterprise AI audit gap that organizations face regarding persistent credentials and ambiguous ownership.