Boards are becoming increasingly focused on understanding the mechanics and implications of agentic artificial intelligence, but traditional governance processes aren’t built for the speed and complexity of today’s AI-driven innovation cycles, said JoAnn Stonier, former chief data and AI officer and chief privacy officer at Mastercard.
Stonier, now working as a consultant focused on data, AI and privacy strategy, said the learning curve is steep, and board directors often lack safe spaces to ask foundational questions about how AI agents are designed, deployed and governed. Without that baseline knowledge, boards struggle to probe issues such as model guardrails, data lineage, incident response and employee impact.
“What we were trying to encourage was board members having a baseline understanding of how agents are created and the difference between generative AI and agentic AI,” Stonier said.
A stronger education foundation, she said, enables executives and directors to assess ROI beyond traditional efficiency metrics. Agentic AI’s real value often emerges in unexpected ways that require a new way of evaluating success metrics beyond cost savings.
In this audio interview with Information Security Media Group at the AI Summit in Manhattan (see audio link below photo), Stonier also discussed:
- Why boards need deeper visibility into the architectural choices behind AI systems;
- How regulated industries can innovate responsibly without adopting a “break things” mindset;
- Why AI spending needs to shift from centralized tech budgets to cross-functional investment models.
Stonier is president at The Cantellus Group, where she focuses on data, AI and privacy strategy. A recent Fellow of Data and Artificial Intelligence at Mastercard, she was also the company’s first chief data officer and first chief privacy officer.