In light of the increasing reliance on artificial intelligence (AI) systems, three additional vendors have recognized a pressing need for tools that can address and reverse AI-related mistakes. This follows the collaboration between Cohesity, ServiceNow, and Datadog, aimed at providing a recoverability service designed to identify and restore files and data affected by malfunctioning AI components to a “trusted state.”
These three companies believe that enterprises will increasingly deploy autonomous AI, also known as “AIOps,” to manage their operations. However, they also anticipate that such software may sometimes produce errors or become targets of malicious attacks. As a result, they argue that organizations will require effective recovery tools to mitigate any issues arising from AI errors.
While users have the option to wait for the market to develop better, more reliable AI technologies less prone to errors and attacks, vendors are complicating this cautious approach. Many are integrating autonomous automation into their products, equipped with tools that not only identify problems but also propose solutions.
Cohesity has announced its intention to launch this product by the end of the year. Upon its debut, Cohesity will compete with Rubrik, which unveiled a comparable tool in August 2025, along with native rollback functionalities already incorporated into solutions from companies like Cisco.
Market analysts predict that this sector has room for multiple players. Gartner estimates that by 2026, up to 40 percent of enterprise applications will feature integrated task-specific AI agents, a substantial increase from less than five percent in 2025.
Meanwhile, Forrester has cautioned that addressing agentic AI challenges will require developers to implement stringent safeguards, including guardrails, identity and access management protocols, and robust oversight.
Cohesity asserts that it mitigates the risks associated with AI agents by creating immutable snapshots of AI environments, allowing for recovery at specific points in time for agents, data, and essential infrastructure. This includes files, databases, object storage, SaaS applications, vector stores, and agent memory.
With the support of ServiceNow and Datadog, which provide comprehensive control and observability platforms, Cohesity enhances its ability to monitor for anomalies. If any issues arise, Cohesity’s tools can initiate API-driven restoration processes throughout the IT ecosystem, ensuring the recovery of AI agents, agent memory, vector databases, model configurations, training and tuning data, and enterprise data repositories.
