Cohesity extends resilience to AI agents, plans to automate cyber resilience

Cohesity announced Agent Resilience (CAR) at its Catalyst event, a product that discovers, protects, and recovers the infrastructure behind enterprise AI agents using the same snapshot architecture, immutable backups, and clean-room recovery as Cohesity Data Cloud. CAR protects agent memory and configuration with point-in-time recovery and also protects the databases, file systems, and services agents interact with. Cohesity Chief Product Officer Vasu Murthy said, "Detection can tell you an AI agent went off course. It cannot undo the changes," and the company outlined a vision for Autonomous Cyber Resilience that uses agentic workflows to automate its five-step cyber resilience framework with a human in the loop.

SECURITY Cohesity extends resilience to AI agents, plans to automate cyber resilience Cohesity has announced Agent Resilience which it says discovers, protects, and recovers the infrastructure behind enterprise AI agents, and has a plan for autonomous cyber resilience, both being announced at its Catalyst https://www.cohesity.com/catalyst/ event. Cohesity Agent Resilience CAR protects agent memory and configuration using the same snapshot architecture, immutable backups, and clean-room recovery that Cohesity Data Cloud customers use to protect workloads and data across on-premises, cloud, and SaaS environments. It has two basic functions, one being to protect agent state, meaning agent memory and configuration, with point-in-time recovery capabilities to restore agents to a known-good state following memory corruption, misconfiguration, or malicious activity. CAR also protects databases, file systems, and other services agents interact with, what agents manage, enabling precise recovery of affected resources when needed. Vasu Murthy, Chief Product Officer at Cohesity, said: “Detection can tell you an AI agent went off course. It cannot undo the changes. Cohesity Agent Resilience provides a precise recovery path for both the agents and the resources they impact. Our vision for Autonomous Cyber Resilience is to use agentic workflows to reduce manual work from response and recovery.” Cohesity is viewing agents as applications that can sometimes go wrong and need restoring to clean state. Enterprises currently have hundreds to a thousand or more agents, and agent protection and resilience is going to grow in importance. Agent topology provides a unified view of each agent; its memory stores, connected applications, databases, and supporting infrastructure. This makes it easier to understand agent dependencies, assess protection coverage, and identify the resources required to restore trusted operations during incidents. Autonomous cyber resilience Cohesity says it has a vision for Autonomous Cyber Resilience, which uses agentic workflows to automate its five-step cyber resilience framework: 1. Protect data, identity, applications, and agents, 2. Help ensure recoverability in all scenarios, 3. Remediate cyber and AI threats, 4. Practice application recovery, 5. Optimize data and AI risk posture. Its intention is to move beyond manually configured, point-in-time plans toward a system that continuously discovers and protects data, validates recovery readiness, and adapts response and recovery plans as conditions change, with a human in the loop. As attackers increasingly use AI to move from initial compromise to operational disruption, resilience plans that rely on manual triage and sequential approvals most probably will not keep pace. Murthy said: “Cyber resilience is not a one-time assessment. It requires teams to protect data, test recovery, detect threats, rehearse responses, and continually improve their risk posture. Agentic workflows can take on more of that routine work and help keep the plan current.” Security or resilience teams would define objectives through Cohesity Copilot rather than manually configuring policies across applications. Eg, all applications and data required to sustain critical business operations, like a minimum viable business, during a cyber incident must meet specified recovery time objectives RTOs and recovery point objectives RPOs and adhere to specified threat scanning and recovery testing requirements. Cohesity’s Data Cloud https://www.blocksandfiles.com/ai-ml/2026/06/18/cohesity-goes-agentic-with-headless-protection/5258151 would then assess the resilience posture of those workloads, recommend appropriate protection policies, threat-scanning strategies, and recovery-rehearsal plans, and present those recommendations for approval before execution. During a cyber incident, admin staff could initiate automated response and recovery workflows to assess the scope of impact, identify indicators of attacker activity, and prepare an isolated recovery environment or clean room to help investigation and recovery efforts. This Autonomous Cyber Resilience model builds on Cohesity RecoveryAgent https://www.blocksandfiles.com/ai-ml/2025/09/18/cohesity-is-protecting-identities-and-active-directory/1595153 , which orchestrates key parts of incident response and recovery and helps security and IT teams work from the same plan. Cohesity Maestro, with expanded capabilities expected later in 2026, will connect Cohesity’s protection, response, and recovery capabilities to customers’ preferred AI tools, including Claude, ChatGPT, and Gemini, as well as Cohesity Helios, the company’s unified management console. Cohesity plans to add more automation as the Autonomous Cyber Resilience vision moves toward commercial availability. AI Resilience Academy The company is laying on AI education courses through its AI Resilience Academy. A Its first course, Foundations of AI Resilience with Cohesity, starts with a simple distinction: AI assistants make recommendations; AI agents can make changes. The free, self-paced course takes 25 to 30 minutes and focuses on concepts and vocabulary rather than product configuration. The course is a free entry point into the Cohesity Academy. Cohesity may offer more advanced AI Resilience courses for a fee over time. Availability Cohesity Agent Resilience is a Cohesity Data Cloud capability and supports Amazon Bedrock AgentCore and Bedrock Agents. It’s available now to select customers with general availability targeted for the end of 2026. Support for Azure and Google platforms is on its roadmap. Automated protection for newly discovered sensitive data is now available to customers with Cohesity Data Cloud Enterprise Edition and Cohesity DSPM. Find out more about Cohesity Agent Resilience here https://www.cohesity.com/blogs/cohesity-agent-resilience/ , about Cohesity's vision for Autonomous Cyber Resilience here https://www.cohesity.com/blogs/autonomous-cyber-resilience/ , and the Cohesity AI Resilience Academy here https://www.cohesity.com/academy/ . Bootnote 1 In July, Druva announced https://www.blocksandfiles.com/data-protection/2026/07/21/druva-spreading-ai-agent-cyber-resilience-wings-with-claude-and-copilot-co-operation/5275619 it was developing protection for AI agents, backup data access for AI agents, protection against AI agents, and better cyber-resilience investigations with AI agents; a 4-way AI agent street. It was extending its SaaS data protection to include Claude and Copilot, letting them access Druva assets via MCP, adding a root cause diagnostic agent and having backups auto-lock when an AI attack is detected. Rubrik's AgentCloud https://www.blocksandfiles.com/ai-ml/2026/06/10/rubrik-using-ai-for-a-service-interface-anthropic-claude-safety-net-and-cloud-app-stack-recovery/5253408 is an agent operations and governance suite, which uses technology from its acquisition of Predibase https://blocksandfiles.com/2025/06/26/rubrik-buying-predibase/ . It enables customers to deploy Claude-powered agents at scale with observability, control, agent rewind to reverse unintended actions, and immutable codebase recovery when an incorrect action outruns version control. A layer of resilience for Claude agents backs up, monitors, and restores the configuration that governs how agents behave. Commvault's AI Protect https://www.blocksandfiles.com/ai-ml/2026/04/14/commvault-announces-trio-of-ai-tools-for-data-prep-agent-oversight-resilience/5217385 , announced in April, will discover and inventory agents across environments and map their activity to AI stacks. It will be designed to discover AI agents and their dependencies operating across connected environments on a recurring basis, helping maintain a unified, up-to-date inventory based on configurable discovery cadence. Each agent record will capture its execution environment and the data sources, models, configurations, applications, and infrastructure with which it interacts. It will help provide a complete, cross-environment picture of what’s running and what it touches. Backup and resilience vendors that deploy agents themselves will readily understand that agents can make mistakes and need restoring to clean states. Cohesity, Commvault, Druva and Rubrik are each determined to provide AI agent protection and resilience. Bootnote 2 Cohesity released its fifth annual Global Cyber Resilience Report https://www.cohesity.com/dm/global-cyber-resilience-report/ , finding that 78 percent of organizations focus cyber recovery efforts on restoring systems rather than maintaining business operations. Restoring systems alone, however, does not guarantee a business can resume normal operations. Recovery depends on more, including whether restored environments can be verified as clean and safe, applications and dependencies are functioning properly, and employees have reliable access to systems and data.