putilov_denis - stock.adobe.com

Cohesity Gaia enables GenAI content creation using backups

A new generative AI capability in the Cohesity Data Cloud platform enables content creation from backup data and aims to eliminate data warehouses for common enterprise uses.

Cohesity's latest generative AI product aims to transform backup data into a data lake that users can query and utilize to create written content.

Cohesity Gaia is a new conversational GenAI assistant that's been added to the Cohesity Data Cloud (CDC) platform. Gaia, which stands for general AI application, indexes customer data stored in the CDC through a supported large language model (LLM) to create reports, audits, knowledge bases and other written content. Security features such as user and data access controls are included.

Many of Gaia's capabilities are available in data warehouse or lakehouse software from the likes of Snowflake and Databricks, according to Brent Ellis, an analyst at Forrester Research. Cohesity Gaia's integration with customer backups and its SaaS subscription model eliminates some technical hurdles of setup and the cost of additional storage for a data lake.

"They've made it so an SMB could leverage [GenAI] on their own data," Ellis said. "This is the easy button version of [a data warehouse]. This is for enterprises that don't want to spend their entire budget on GPUs or the cloud [for GenAI]."

A list of backup vendor's GenAI offerings.
Cohesity's Gaia is the latest in a string of integrations from backup vendors. This list was compiled at the end of 2023.

Gigante data management

Comparing Gaia with data warehouses wasn't made by market analysts alone. Sanjay Poonen, CEO of Cohesity, also talked of Gaia as being akin to a Snowflake or Databricks -- although with added security.

"[We're] a data and security company that's using AI to deliver insights and keep the data safe," Poonen said.

Cohesity previously classified its GenAI programs under the brand name Turing but now uses that name to encompass all GenAI initiatives and products. Specific releases will still bear unique titles, such as Gaia.

Gaia queries against backup data in the CDC designated by the user, according to Cohesity, and includes role-based access controls for compliance. Backup data indexed by Gaia still maintains its original security features, such as immutability.

The Gaia service will eventually integrate with other Cohesity tools, such as its DataHawk threat detection service, according to the vendor. This integration could provide further insights such as specific points of corruption or security weaknesses in plain English.

The Gaia capability is available now as a SaaS and priced according to the total storage indexed by Gaia itself and the number of customer queries. ChatGPT-3.5 and ChatGPT-4 LLMs are supported, and a free trial is available. Support for additional LLMs and user-created models is planned for future versions, according to Cohesity, along with support for an on-premises version.

Backup titans to rise

Other backup vendors are integrating GenAI capabilities into their suites. But the specific capabilities are still in significant flux during these early days of enterprise adoption, according to backup market analysts.

Competitors such as Commvault's Metallic or Druva's Dru bring GenAI capabilities to different aspects of the backup administrator's job, Ellis said. Metallic introduced Arlie in November, but the GenAI chatbot is limited to diagnosing security concerns compared to the more general content creation of Gaia.

[We're] a data and security company that's using AI to deliver insights and keep the data safe.
Sanjay PoonenCEO, Cohesity

Those generalist capabilities might be singular to Cohesity now, but Ellis said he expects other vendors to catch up quickly, even if the market itself is slightly cold to having GenAI in backup software.

"This is not unique to Cohesity. But what they're doing is bringing a fuller vision of [GenAI services] to the market," Ellis said. "I don't know if this will resonate with the market yet."

Using backups for GenAI content is still a unique use of an asset many enterprises already have, according to Krista Macomber, research director at The Futurum Group. Implementing LLMs or GenAI capabilities is challenging, but using existing data for such projects could have enterprise value.

"The thing to keep in mind is that issues around optimizing performance, bandwidth and costs also apply in the secondary storage environment as well," Macomber said. "It's very complex, time-consuming and can be expensive to implement. This is really the key issue that Cohesity is looking to address with Gaia."

Tim McCarthy is a journalist from the Merrimack Valley of Massachusetts. He covers cloud and data storage news.

Dig Deeper on Data backup and recovery software

Disaster Recovery
Storage
ITChannel
Close