From Demos to Deployments: How U.S. Bank Is Industrializing AI

Proving AI works was never the hard part. The hard part is industrializing it. In this episode of Technovation, Peter High speaks with Dilip Venkatachari, Chief Information and Technology Officer at U.S. Bank, about what it actually takes to move from scattered AI experiments to a common platform with reusable components and measurable outcomes. Drawing from eight years leading technology at one of the country’s largest banks, Venkatachari explains how U.S. Bank treats model risk management and regulatory requirements as design parameters rather than obstacles, why the real technical challenge is plumbing rather than models, and how an internal AI marketplace uses social incentives to drive reuse without mandating it.

Key Highlights:

Why industrialization, not proof-of-concept, is the defining AI challenge for large enterprises

How U.S. Bank treats regulatory requirements as design inputs rather than blockers to AI deployment

Why the best AI system creates no value if people don’t change their workflows

How an internal AI marketplace and citation-style reuse incentives prevent redundant technology stacks

Why AI is emerging as an unexpected solution to the COBOL skills retirement crisis

How frontier model companies are shifting from LLM providers to enterprise deployment partners

This episode is sponsored by Tines. This episode is presented by Tines: Freedom to build. Complete security. No compromises. Learn more at tines.io