Governing Agentic AI at Enterprise Speed

A conversation on what the agentic workforce demands from enterprise architecture, and why enforcement must start at the data layer.

Listen to the episode

Also available on:

The governance models enterprises built over the last decade assume a human in the loop—someone to consult policy and carry accountability for what follows. In the agentic AI era, that assumption breaks.

Priyanka Jain, EDB VP of Product Management for Data & AI Governance, calls what's at stake now the "governed velocity of innovation": the ability to deploy AI fast precisely because governance is already embedded in the architecture.

That governance has to start at the data layer, where agents actually touch data. Permissions need to be verified and actions reconstructed at the point of execution, at agent speed. EDB CLO Rob Feldman's concept of the "digital leash" makes this concrete: a system that enforces what agents are allowed to do in real time and creates a feedback loop when something goes wrong.

Together, Jain and Feldman map out a phased approach: read-only, low-risk deployments first to prove out the foundation, then expanding scope as enforcement earns trust. When governance is built into the data layer from the start, every subsequent deployment gets faster.

The organizations that build this architecture now are laying the foundation for the agentic enterprise.

Key takeaways:

Agents break the governance model. Governance was built for predictable actors. Agents are non-deterministic, executing without policy consultation at speeds existing review cycles were never designed to handle.
Enforcement has to live at the data layer. Policies defined at the business layer have no mechanism to stop an agent in motion. The execution boundary is the only place governance can actually hold.
Governance can't wait for regulation. Things are moving too fast, and customers will withdraw trust from organizations running ungoverned AI long before any legal framework catches up. The governance question is already a market question.
Governed velocity is an architecture decision. Organizations that move enforcement to the data layer from the start get faster with every deployment. The foundation does the work that review cycles used to do.

About the guests

Priyanka Jain, VP, Product Management – Data & AI Governance, EDB

Priyanka is responsible for the product, data and AI governance functions at EDB, where she is defining how enterprises govern agentic AI at the data layer. A seasoned product executive, she has spent her career at the intersection of data, analytics and regulation, building the platforms that let organizations adopt advanced technology without surrendering control or trust. Priyanka spent almost a decade at IBM leading data management and advanced analytics, and she went on to build data and AI platforms for some of the most heavily regulated industries in the world, including healthcare and pharmaceuticals.

Rob Feldman, Chief Legal Officer, EDB

Rob is responsible for the worldwide legal and compliance functions at EDB, including EDB's Responsible AI initiatives. An experienced executive and lawyer, he builds high-performing legal teams to support growing technology companies in dynamic business and regulatory environments. Rob spent more than a decade in private practice as a technology company litigator, focused in securities fraud defense, intellectual property disputes and government and internal investigations. Rob also serves on the UN Global Compact Legal Council, providing strategic guidance on global regulatory environments to help businesses drive transformative, long-term impact.