Cohesity Secures Industry-First Patent for RAG Platform Leveraging Secondary Data

Cohesity, a leader in AI-powered data security, announced that the United States Patent and Trademark Office (USPTO) has granted the company patent No. 12,619,501 for the underlying technology of its enterprise generative AI platform, 'Cohesity Gaia.'

The patent, titled 'Data Retrieval Using Embeddings for Data in Backup Systems,' covers Cohesity's unique approach to enhancing generative AI applications by combining secondary data systems with a Retrieval-Augmented Generation (RAG) semantic layer. Cohesity is the first data protection vendor to receive a patent for this specific approach.

This technology enables customers to apply generative AI directly to secondary data without creating data silos, weakening governance, or increasing the risk of sensitive data exposure. It establishes a security-first framework that allows AI workloads to run directly on protected secondary data while maintaining existing security, governance, compliance, and access controls.

The inventors are Sanjay Poonen, Gregory Statton, Mohit Aron, and Apurv Gupta. This serves as a core innovation for 'Cohesity Gaia,' turning secondary data—historically used for recovery and maintenance—into a secure knowledge source for generative AI.

'Protected data is a treasure trove; it's the most important and complete repository of corporate knowledge,' said Sanjay Poonen, CEO of Cohesity. 'This patent is the result of years of foundational engineering to change the status quo with a security-first architecture. We allow customers to maximize insights without forcing them to move or replicate sensitive data.'

Cohesity Gaia is currently available as part of the Cohesity Data Cloud platform, assisting teams in accelerating decision-making and building AI-powered workflows.