Commercial loans can take weeks to close, and Pennsylvania-based Customers Bank put its typical wait at 30 to 60 days. A pilot of its artificial intelligence (AI)-powered process closed select commercial loans in seven days, the bank said in July in its Q2 earnings release.
Customers Bancorp, the bank’s parent with more than $27 billion in assets, said Thursday (Oct. 22) for its Q3 earnings results that it is scaling the process. It also doubled its deployed AI agents in the quarter, a count that includes custom GPTs and similar tools.
Every AI use case at the bank follows one cycle, Customers Bancorp CEO Sam Sidhu said on the bank’s July earnings call. Engineers pair with the bankers who own the work and shadow the real workflow. They build agents on the bank’s own data and systems, then measure the result before the agent joins daily operations. OpenAI engineers sit beside the team to build custom tools for the bank’s processes, Sidhu said. He added that the tools are proprietary and built in-house.
Sidhu personally leads a 40-person team, about 5% of the workforce, focused on AI workflow transformation. In April, when Customers Bank announced its work with OpenAI, Sidhu said bankers would “spend more of their time on the work that creates value” by the end of 2026, PYMNTS reported.
Multi-Agent Underwriting Cuts Time to Close 85%
Customers Bank engineers built a multi-agent credit underwriting process, Sidhu said. In a second-quarter pilot, the process closed commercial and industrial loans and commercial real estate loans within a week. Sidhu cited an 85% cut in the time to “readiness to close,” compared with industry norms of 30 to 60 days.
We’d love to be your preferred source for news.
Please add us to your preferred sources list so our news, data and interviews show up in your feed. Thanks!
Customers Bank finished the pilot one to two quarters ahead of schedule, its July release said. Larger builds still take quarters, not weeks, and the underlying technology is about six months old, Sidhu said.
Customers Bancorp listed broad deployment of its seven-day loan closing tool among its 2026 AI priorities in July. On Thursaday, it said the loan process added functionality and drew more use in the third quarter.
Banks Count AI Gains in Dollars, Hours and Ratios
Customers Bank employees said in a survey that they saved an estimated 76,000 cumulative hours, up nearly 65% from the second quarter. Customers Bancorp also estimated more than $15 million in annualized capacity, which it counts as salary and benefits savings from positions not backfilled and open positions reduced this year.
Full payoffs are still rare. Among large U.S. companies in financial services, healthcare and media, between 90% and 100% report some return on new AI tools. Only 5% to 10% report a full return, according to the PYMNTS Intelligence report “The Enterprise AI Payback Curve: Adoption Accelerates as Returns Take Shape.” The June 2026 survey covered 60 senior technology executives at companies with at least $1 billion in revenue, so the figures are self-assessments.
Bigger lenders are putting their own figures on AI. TD Bank Group, the Canadian lender, delivered about 195 million Canadian dollars (about $141 million) in AI value in the first three quarters of its fiscal year, against a full-year target of about 200 million, PYMNTS reported.
Bank of America CFO Alastair Borthwick said in July that AI capabilities contributed to a 59% efficiency ratio, about 360 basis points better than a year earlier, Fortune reported.
Lloyds Banking Group, the U.K. lender, has more than 800 AI models live and targets a cost-to-income ratio below 45% by 2030, PYMNTS reported in July.
Leave a Reply