Wells Fargo's Faraz Shafiq: four keys to AI-led bank overhaul

Wells Fargo's AI effort is being shaped by a leader who joined in February from Amazon Web Services to oversee the roadmap and development of AI-powered products across the company. Faraz Shafiq's approach reflects a shift from enabling tools for employees at the start of the year to identifying targeted uses that are now showing tangible benefits.

The bank's hub-and-spoke model places governance, security, controls and prioritization in one central team, while delivery teams with domain expertise own specific applications. Shafiq argued that a small number of use cases will create a disproportionate share of value, so Wells is concentrating effort rather than stretching thin. On pace, he said the bank wants a faster operating model but not at the expense of safety and guardrails, framing the conversation as "what do we need to do to do this safely, effectively?" rather than "we can't do that."

Wells is also pointing to early results: virtual assistant Fargo has gathered data through some 1 billion customer interactions, banker intelligence capabilities are lifting product sales rates by 40%, and a small language model built by the model risk management team outperforms off-the-shelf models for debit card fraud detection. The bank's move from No. 6 to No. 5 in this year's Evident AI index was driven largely by an 11-point increase in leadership, and Wells last week said it hired Cayetano Gea-Carrasco as head of AI and data for consumer banking and lending.