U.S. Bank is entering the next phase of artificial intelligence with a clear test for progress: the first use case may require a village, but the tenth should take an afternoon. That principle captures the shift the company’s Dilip Venkatachari is leading from proving that AI works to industrializing it across one of the country’s largest financial institutions.

U.S. Bancorp, the parent company of U.S. Bank, reported record net revenue of $28.7 billion in 2025. The Minneapolis-based company serves millions of customers through consumer banking, business banking, commercial banking, institutional banking, payments and wealth management. Its payments businesses also support roughly 1,000 other financial institutions. As senior executive vice president and chief information and technology officer, Venkatachari leads global technology infrastructure, security, products and services as well as digital and technology transformation.

From Proving AI to Industrializing It

U.S. Bank’s early AI efforts focused on establishing that the technology could create customer value within the controls required of a bank. Venkatachari now describes the institution as moving from a “prove-it phase” to a “scale-it phase.” “The hard part here isn’t proving that AI works,” he noted. “The hard part is industrializing it.” That requires a common foundation, strong governance, reusable capabilities and a model that translates use cases into measurable results. Security, privacy and regulatory obligations are design requirements that must scale alongside innovation.

The greatest obstacles are often less glamorous than the models. “The challenge for us from a technology standpoint are not the models. It’s the plumbing,” Venkatachari emphasized. Data, identity, integration, observability and deployment determine whether an experiment can become a dependable capability. Data must also mean the same thing across businesses and be available for real-time customer interactions.

Reuse Without Smothering Experimentation

U.S. Bank has organized AI as a federated capability. Central teams provide technology foundations, developer tools and governance while AI talent sits close to business teams and customers. The model is intended to preserve local speed without allowing each experiment to create a separate stack.

“We are trying to avoid 1,000 AI experiments becoming 1,000 technology stacks,” Venkatachari explained. The bank has built a common platform with shared services and guardrails, along with an internal AI marketplace where teams can discover and reuse one another’s work. It also uses social incentives inspired by academic citations. When a tool is reused, its creator can see that others value it and may continue improving it.

That approach deliberately favors encouragement over mandates. A rigid reuse requirement can force teams to adopt a capability that does not fit their needs and discourage better ideas. By making reusable components visible and rewarding their creators, U.S. Bank seeks to make each deployment faster and less expensive while retaining room for experimentation.

Changing Work and Preserving the Talent Pipeline

The bank’s workforce agenda extends beyond conventional training because AI tools evolve faster than formal curricula. U.S. Bank teaches enduring principles, runs hackathons and conducts hands-on sessions with frontier model companies and strategic partners, then relies on champions to spread knowledge.

“The answer is not someone who knows everything,” Venkatachari highlighted. “The answer is having people who are comfortable that they’re open, that they want to learn.” Access to AI in familiar tools also helps employees rethink their own workflows before redesigning a team or business process.

Software engineering illustrates the change. Developers are moving from writing every line of code toward directing, reviewing and orchestrating AI-generated work. Venkatachari does not expect that transition to cause an immediate, sweeping reduction in technology employment. Productivity will rise, but demand is also expanding as businesses identify more valuable work. Judgment, domain expertise and customer understanding become more important.

That makes the early-career pipeline a strategic concern. Organizations still need an apprentice model through which employees acquire the context required to evaluate AI output. AI may also ease shortages in older skills by explaining and translating decades-old COBOL applications for newer engineers.

Competing on Time to Value

Venkatachari sees AI becoming easier for enterprises to consume. Frontier model companies are evolving into software solution vendors, middleware is maturing and established SaaS companies are embedding AI into specialized workflows. U.S. Bank is also engaging earlier with startups and investors because promising vendors can reach meaningful bank adoption in months.

That maturation changes the build-versus-buy calculation. Enterprises can now use external scaffolding and specialized products to reach outcomes faster, provided they connect to the bank’s data, controls and applications.

Cost reduction remains part of the opportunity, but Venkatachari argues that speed is the larger prize. AI can help the bank build products, make credit or dispute decisions and personalize services more quickly. “Small improvements in a variety of areas, when multiplied across millions of customers and thousands of employees, can truly become transformational,” he underscored.

The objective is not to own the most advanced model. It is to create an organization capable of converting fast-changing AI capabilities into trusted and measurable impact at scale. “The single most important element for us and for probably every other bank and enterprise is time to value,” Venkatachari said. “What was true for doing that a year ago is not true now.”

Peter High is President of Metis Strategy , a business and IT advisory firm. He has written three bestselling books, including his latest Getting to Nimble . He also moderates the Technovation podcast series and speaks at conferences around the world. Follow him on X @PeterAHigh .