Data security is emerging as the bedrock of enterprise success, mostly out of necessity. As agentic AI scales, faulty data classification and ungoverned access are exposing organizations to serious risk.

But the challenge runs deeper than most organizations realize, according to Ronan Murphy (pictured), chief data strategy officer of Forcepoint LLC. Teams are discovering that overconfident manual data labeling has exposed their flank, with misclassified records now pouring into AI pipelines.

“There’s a common misconception in companies that because they’ve been labeling their data, it’s all correct. But in reality, when you look at it, you’ll find that for the last 15 years, we’ve got a 70% classification [rate] of our data which is accurate — 30% is wrong,” Murphy told theCUBE. “But that 30% finding its way into a copilot or an LLM or an AI application literally means you’re one prompt away from a major data breach, or from losing trade secrets or IP or regulated content.”

Murphy spoke with theCUBE’s Christophe Bertrand at the RSAC 2026 Conference for an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed how data security governance must underpin every AI initiative before agents are given access to enterprise assets at scale. (* Disclosure below.)

Getting the data security right — before agents run free

The current challenge comes down to data at rest, data in motion and data in use — the three main categories that Forcepoint is honing in on, Murphy explained. Each carries its own exposure risk as AI agents access repositories and move information across systems while manipulating documents in real time.

“How do you eat an elephant? One bite at a time,” Murphy said. “What you have to do is you have to look at the data assets that represent the most value to your business. Then you’ll work your way through the respective data assets, make sure you know where they are, make sure they’re classified correctly, make sure they’re in the right location and make sure that they’re aligning with the AI initiative that your organization has.”

With manufacturers, defense contractors, healthcare organizations and retailers all rushing to operationalize AI at the same time, the window for getting the foundations right is shrinking fast, Murphy noted. Companies that skip the data layer are building their AI strategy on shaky ground — while those that get it right will be better positioned to scale AI safely and effectively.

“My overriding view is that companies who get the data layer right will really accelerate their AI journey,” Murphy said. “It’ll become a force multiplier and ultimately demonstrate tremendous ROI and success for that business.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of the RSAC 2026 Conference:

(* Disclosure: Forcepoint sponsored this segment of theCUBE. Neither Forcepoint nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE