The sharp rise of autonomous systems has created a critical need for enterprises to secure AI agent identity with the same rigor used for human employees. But as these digital actors gain access to sensitive systems, who is keeping watch over what they do next?

As digital actors start to handle sensitive transactions and access core systems of record, the risk of unmanaged non-human identities becomes a significant vulnerability for global brands. This is driving a shift toward “just-in-time governance”, where every action taken by an autonomous agent is authorized in real time based on live context rather than standing permissions, according to Andre Durand (pictured), founder and chief executive officer of Ping Identity Ltd., enterprise access security provider. In effect, companies are starting to treat AI agents less like software tools and more like active identities that require continuous supervision.

“The way in which we secure our systems — who has access to what — at its core is all founded on the notion of knowing who is acting,” Durand said. “There is no security without identity. You can’t secure what you can’t identify.”

Durand spoke with theCUBE’s Christophe Bertrand at the RSAC 2026 Conference, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed AI agent identity and the transition to runtime decision-making for autonomous actors. (* Disclosure below.)

Strengthening guardrails for the AI agent identity channel

Unlike human workers, digital agents lack personal accountability, making the implementation of zero-trust principles critical for enterprise systems. While companies have spent years on a journey toward zero trust for humans, trust in the agentic world can break down much faster, Durand explained.

“With agents acting in our system, there’s no equivalent of a consequence for something that we deem to be … damaging to the company,” he said. “For that reason, the guardrails need to be much tighter.”

To address this, Ping Identity launched its Identity for AI platform, aimed at managing AI agents from registration and delegated access to runtime enforcement. The offering is designed to improve visibility into agents across environments and ensure they operate with only the access needed for a given task, according to Durand.

“While we’re all compelled to move fast, and there’s incredible competitive advantage to the level of efficiency and scale that we can get out of the agentic landscape, it’s not to be done carelessly,” Durand said. “This is where all the security principles and the right foundation … need to be harnessed appropriately to make sure that we don’t have rogue things that are unexplainable impacting our brand and our companies.”

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

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

Photo: SiliconANGLE