Enterprise AI adoption often outpaces the governance built to support it, leaving controls retrofitted onto tools in production. Beth Hiatt, SVP of Product Management, Data Governance and Responsible AI at LPL Financial, opens a three-part conversation with Jessie Smith, Chief Product Officer at Ataccama.

Hiatt draws on more than 25 years of experience in data management and governance to describe why AI programs need a blueprint before organizations expand across multiple use cases. That blueprint includes data segregation, semantic layers, knowledge graphs, catalogs, business meaning, model management and a control plane.

The conversation also examines how governance can become part of daily execution. Hiatt argues that governance should be embedded in software development life cycles and risk assessment processes, with governance teams working with teams developing AI use cases.

Hiatt also flags the first 90 days as the window to build the right infrastructure. Four themes run through the rest of the conversation.

• Running both postures at once: Data programs used to pick a defensive or offensive starting point, but AI governance has to hold both at the same time.
• The ninety-day blueprint: Before scaling AI use cases, Hiatt says the underlying data infrastructure has to be mapped out first.
• Governance on the ground: Moving past a reputation for saying no means governance becomes part of the process instead of a policy layered on top.
• Context over metadata: Classification and lineage now track whether data can train a model, since AI without business context will not produce the results a company expects.

On the move? Take this conversation with you. If you don’t have time to finish the video, you can stream the complete, uninterrupted audio edition of this episode on the go.