Urgent.News

What's breaking now, across thousands of outlets.

Tech

Agents want the crown jewels: HPE and Daxa push governance down to the data layer

Enterprises are handing AI agents their most sensitive data, and governance is struggling to keep pace. Techniques such as semantic caching are emerging as a way to enforce controls inline without slowing agents down. Moving agents from pilot to production means letting them read, change and even delete the records that matter most. That shift […] The post Agents want the crown jewels: HPE and…

Agents want the crown jewels: HPE and Daxa push governance down to the data layer

Agentic AI agents are increasingly given access to an enterprise's most sensitive data, prompting a surge in the need for robust governance. Techniques such as semantic caching are being developed to enforce security controls inline, without slowing down the agents. As these agents move from pilot to production, they require the ability to read, modify, and even delete critical records, making data access the primary security concern.

Huseni Saboowala, co-founder and CEO of Daxa Inc., emphasized that agents must have the capability to interact with real enterprise data, including intellectual property and customer information, to truly deliver value. This expanded access, however, significantly increases the risk of data manipulation. While traditional security measures like firewalls, data loss prevention, and identity management remain essential, they are designed for users and applications, not data-centric AI agents.

Saboowala explained that these agents operate at a cognitive layer, where prompts, responses, retrievals, and tool calls occur, a layer where traditional security tools are ineffective. To address this, Daxa is partnering with Hewlett Packard Enterprise (HPE) through their Unleash AI program, which provides pre-validated software on HPE Private Cloud AI.

Braun, HPE's vice president of AI business development, highlighted that incorporating governance into the AI stack from the outset allows for better understanding and integration, rather than attempting to add it later. Timing is crucial, as offline governance is too late for agents working at machine speed. Inline enforcement through semantic caching, small language models, and a distributed data plane is necessary to ensure security.

A regional bank that implemented Daxa's solution in conjunction with HPE has already connected its agents to sensitive source code and customer data. Saboowala concluded that the future of successful AI agents lies with enterprises that can trust their data with these agents.

Written by urgent.news from SiliconANGLE's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at siliconangle.com →

More in Tech

More from Wednesday 18 November →