Cohesity has introduced Cohesity Agent Resilience, a new capability designed to discover, protect and recover the infrastructure behind enterprise AI agents, as organisations grapple with the risks of autonomous systems operating with limited human oversight.

Announced at the company’s Cohesity Catalyst event, the launch is aimed at closing what Cohesity describes as a critical gap: while governance and observability tools can flag when an AI agent behaves unexpectedly, they cannot restore data or systems once an agent has altered, corrupted or deleted them.

Protecting AI agents and what they manage

Cohesity Agent Resilience protects agent memory and configuration using the same snapshot architecture, immutable backups and clean-room recovery that the company’s customers already rely on for sensitive workloads. It also protects the databases, file systems and other services that agents interact with, enabling more precise recovery when something goes wrong.

At launch, the capability supports Amazon Bedrock AgentCore and Amazon Bedrock Agents, with support for Microsoft and Google agent platforms on the roadmap. It is available now to select customers, with general availability targeted for the end of 2026.

The launch follows new findings from Cohesity’s fifth annual Global Cyber Resilience Report, conducted with Vanson Bourne, which found that 56 per cent of organisations felt unprepared to detect or contain unintended actions by AI agents and automated workflows, while 58 per cent lacked confidence in their ability to verify the integrity of AI models and related data after a cyberattack.

“Detection can tell you an AI agent went off course. It cannot undo the changes,” said Vasu Murthy, Chief Product Officer, Cohesity. “Cohesity Agent Resilience provides a precise recovery path for both the agents and the resources they impact.”

A vision for autonomous cyber resilience

Alongside the new capability, Cohesity outlined its broader vision for what it calls Autonomous Cyber Resilience — using agentic workflows to automate its five-step cyber resilience framework of protecting data, identity, applications and agents; ensuring recoverability; remediating threats; testing recovery; and continuously optimising risk posture.

“As AI agents take on more business-critical work, we need to think beyond monitoring and make sure we can recover when something goes wrong,” said Srikanth Kuntamukkala, CIS Global Lead for Partner Sales, Cognizant.

Cohesity also launched the Cohesity AI Resilience Academy, beginning with a free, self-paced course aimed at helping organisations understand the risks of AI systems moving from advising people to taking autonomous action.