Cohesity Inc. today announced that customers can request early access to Cohesity Turing, a set of artificial intelligence features it first previewed in May. 

To support the AI features’ rollout, the data management provider is launching a new technical collaboration with Amazon Web Services Inc. The collaboration will see the companies integrate Turing with the cloud giant’s Amazon Bedrock generative AI service. Cohesity previously introduced a similar integration for Google LLC’s Vertex AI offering, which competes with Bedrock.

“For some time, enterprise IT priorities have focused on managing data proliferation, data security, and compliance,” said Cohesity Chief Executive Officer Sanjay Poonen. “We now see a rapidly increasing demand for AI-powered data insights from customers.”

San Jose, California-based Cohesity provides a data management platform called the Cohesity Data Cloud. It enables organizations to back up business files, reduce the amount of storage space their information takes up and perform a range of related tasks. As of late 2021, the platform was used by more than 2,500 organizations including nearly a quarter of the Fortune 500. 

Turing, the AI feature suite that Cohesity is making available in early access, debuted six months ago. It allows administrators to search the data their organizations store in Cohesity’s platform using natural language commands. Turing lends itself to more advanced tasks as well: an administrator could, for example, use it to analyze employees’ data access patterns for signs of unauthorized file usage.

Cohesity says that the feature suite can also help companies implement RAG, or retrieval augmented generation. This is a machine learning technique enterprise developers are applying in a growing number of AI projects.

The dataset that a neural network draws on to answer user queries can become outdated over time. Historically, refreshing a neural network’s dataset required retraining it, which is an expensive and time-consuming process. RAG, the machine learning technique that Turing can help companies implement, makes it possible to refresh or extend an AI model’s knowledge base without retraining it.

Cohesity will use AWS’ Bedrock service to power some of Turing’s capabilities. Bedrock provides access to a collection of managed AI models hosted on the Amazon Inc. unit’s cloud platform. Developers can access those AI models through an application programming interface without having to manage the underlying infrastructure.

Cohesity detailed the collaboration alongside a number of upcoming features it plans to introduce at AWS re:Invent 2023 next month. DataProtect, a tool in the company’s data management platform that eases file backup, recovery and related tasks, is receiving support for Amazon S3. Additionally, Cohesity is enhancing the tool’s existing integration with the Amazon RDS relational database service. 

Photo: Cohesity