Wells Fargo CEO Charlie Scharf is pushing back on the idea that artificial intelligence is simply good or bad for employment, telling investors the reality inside his bank is messier than either camp admits, according to a published report.
Speaking at an investor conference last week, Scharf said it is “obvious” that AI is both a threat to jobs and a creator of them, according to the report. The complication, he said, is timing and fit: the people displaced are not always the people the technology creates roles for.
Scharf pointed to several functions where the bank sees room for AI to streamline work, including auditing, testing, legal, contracts, and credit memos, per Banking Dive‘s account. Those are the same back-office and compliance-heavy areas many ARM operations are now eyeing for automation. He cautioned that efficiency gains may not all fall to the bottom line, since competitors are chasing the same tools, but called the effect a net positive for the bank’s expense base.
Notably, Scharf said the bank still plans to hire, both people who can build models and staff who can use the new tools to serve customers better. “People don’t go away in this,” he said. The bigger problem, in his view, is a workforce “mismatch” the country will have to manage, which Wells intends to address partly through retraining.
The numbers behind the strategy are striking. Banking Dive reported that headcount has fallen from roughly 275,000 when Scharf took over in late 2019 to about 201,000 as of the first quarter, and that the bank has stripped out around $15 billion in expenses while adding roughly $9 billion in new investments. Wells also recently emerged from the asset cap tied to its 2016 fake-accounts scandal, lifted in June 2025 after seven years.
The takeaway for collection and recovery leaders: the largest creditors are treating AI as a tool for margin and headcount discipline at the same time, not one or the other. Vendors and agencies that can show measurable efficiency, without adding compliance risk, are likely to find a more receptive audience among bank clients running the same math.
