Vinod Shenoy

Everyone Wants AI, No One Wants the Accountability but how do you Rethink Governance for the Agentic Era? As organisations rush to experiment with generative and agentic AI, many are discovering that their mature datam governance frameworks are not enough. AI introduces a different class of risk, accountability, and lifecycle challenges: it is no longer just about whether the data is good, but whether the entire AI solution continues to behave as intended—five days, five months, or five years after deployment. In this session, unpack the critical differences between data governance and AI governance, and why treating them as the same discipline leads to ownership gaps, blurred accountability, and failed scaling of promising pilots. Drawing on practical experience from building and governing AI solutions in a highly regulated environment, explore:

• Designing governance around evaluation, measurement, and continuous monitoring of AI outcomes

• Where AI governance should sit organisationally – and what happens when "everyone wants the AI but no one wants the accountability"

• Moving beyond pilots by addressing integration, tooling, and siloed experimentation

• Pragmatic steps CDOs can take to evolve from data governance to AI governance that the business can trust