Saul Van Beurden has a folder on his phone labeled “Productivity.” Even as he sits in his 60th-floor office at Hudson Yards, he constantly references the many AI apps that exist there. Or in front of his broken washing machine.
Van Beurden, who was named head of Wells Fargo’s AI division in November, is betting the technology will catapult big retail banks forward after the Federal Reserve lifted its $1.95 trillion asset cap last summer. Some growth has already begun, with the bank planning to open about 3 million new credit card accounts, renovate 700 branches, and make further upgrades. And Wells Fargo has big ambitions, including becoming one of the industry’s “top five” M&A advisors, in part by hiring top talent.
For Van Beurden, AI is essential to achieving these goals and will impact every business area and every workflow, from call centers to investment banking. His job is to keep everything in perspective and prioritize the most urgent and innovative initiatives.
“This creates completely new opportunities,” Van Beurden, a native of the Netherlands, told Business Insider about AI. “If you look at our strategy, it’s very simple: Fundamentally change the way banks operate.”
“Air traffic control”
Banks’ AI infrastructure runs on a “hub and spoke” model. Van Beurden’s small central team will serve as the hub, with designated generated AI leads serving as spokes in each business unit, from commercial banking to human resources. These leads will be responsible for identifying, scaling, funding, and executing use cases. Alongside his AI role, Van Beurden also serves as Co-CEO of Consumer Banking and Lending. He suspects that his dual role makes it easier for leaders on both the business and technology sides of the bank to trust him.
Part of his job is to identify tools that can be used company-wide to discover potential conflicts before they occur.
“If we start to see the possibility of a collision, we will use air traffic control,” he said. “So if two use cases require something at the same time, ROI, or return on investment, determines which one gets done first. This is very clear: the most impactful use case takes precedence.”
While AI agents are becoming central to some high-impact projects, especially those that rely on manual workflows, they are not yet used in common human-to-human interactions.
Wells Fargo has relationships with multiple providers including Microsoft and Google Cloud. Evident AI’s latest tracker of AI maturity in the banking industry ranks the bank in sixth place, ahead of rivals Goldman Sachs, Bank of America and Citigroup.
Recruitment is the key
On any given day, Van Beurden could be seeing how Wells Fargo is faring on its AI journey. His team tracks AI tools and their usage compared to eight other large banks, measuring daily, weekly, and monthly adoption rates. Wells Fargo declined to share the data.
The retail banking giant isn’t mandating the use of AI, in part because Van Beurden believes forcing it can feel like simply “following the boss.” Instead, we want to build “AI literacy” through initiatives such as training, internal demos, and regular communication.
He says leaders must be able to answer basic questions for employees: “Why should we use it? Why is it good? If you can’t answer that question, you’ve failed.”
Three key principles
Van Beurden believes that several key principles help differentiate Wells Fargo: continuous learning, simple messaging, and constant execution.
Our small team is constantly exploring new LLMs, as the model and its features evolve daily and even hourly.
“This is a horse race where different horses are always leading the race,” Van Beurden said of AI vendors. Wells Fargo has built a platform that can react as new models come out, making decisions based on features and cost.
Van Beurden believes in explaining AI to employees as simply as possible, even though he operates in a world full of jargon. He says companies quickly lose talent when they start “hiding behind” technical jargon.
Still, strategies and messages are meaningless unless they are implemented.
“It takes a village to get an AI program off the ground quickly,” Van Beurden said. “Execution is really the most important thing.”
