AI ownership is being settled across the enterprise right now, and the CDO has a narrow window to shape where the lines fall. This session looks at how peers have divided authority over models, budget, and risk with the CIO and with business units running their own initiatives, and what they chose to claim versus concede. Panelists share the operating arrangements that have held, and how they built the case for the remit they wanted.

Most data leaders received an AI mandate without a corresponding increase in budget, making capital allocation the defining skill of the role. This session covers how peers have reprioritized inside their existing envelope, protected foundational investment while funding new capability, and built the business case that unlocked additional spend once early results were in.

Contracts, claims, email, and support records hold significant value and remain largely unusable for AI. This discussion covers how peers decide which corpus to unlock first, what extraction and validation actually cost at volume, and how they prove the return before committing to the next one.

Perfect quality is unaffordable and inconsistent quality is unusable, so the real work is deciding where the threshold sits. Participants compare how they have defined reliability standards by domain, who approves exceptions, and how they moved from periodic checks to continuous monitoring.

Every structure buys speed in one place and pays for it somewhere else. Panelists who have restructured mid-tenure discuss what triggered the change, what it cost in people and delivery momentum, and what they would weigh differently making the call again.