OpenAI partners with Customers Bank in push to automate finance

Nearly half an hour into a conference call on Friday to discuss first-quarter results with analysts, Customers Bank CEO Sam Sidhu revealed something unusual — up until that point, he hadn’t actually been speaking.

“The prepared remarks you heard on my behalf today were delivered by my AI clone, not read by me,” Sidhu said, calling it a potential first for a public company earnings call.

The point of the stunt, he said, was to underscore a broader shift happening as Customers Bank, a $25.9 billion asset lender catering to startups and small businesses, embraces artificial intelligence.

Customers Bank has signed a multiyear partnership with OpenAI in which the AI giant will embed engineers at the company to help it automate lending and client onboarding, CNBC has learned exclusively.

The deal is part of Sidhu’s effort to get ahead of other banks in the industry’s race to transform itself using AI agents as a new digital workforce. His strategy hinges on automating core banking processes — slashing loan timelines from weeks to days, for instance — and scaling growth without adding staff at the same pace.

While many bankers have described AI in broad terms like productivity gains, Sidhu is tying it directly to financial targets.

Sidhu told CNBC that the project will improve the firm’s efficiency ratio from about 49 to the low 40s, boosting the bank’s returns starting next year.

The relationship with OpenAI — which has targeted finance as one of its core industries, even hiring former bankers to train its models — will be a symbiotic one for the AI giant, according to the bank CEO.

“We’re going to be co-creating enterprise solutions they could potentially sell to other banks in the future,” Sidhu said. “The goal here is end-to-end, automated agentic led workflow” for lending, deposits and payments.

OpenAI said it was proud to help Customers Bank “as they build a more intelligent operating model that empowers employees, strengthens client service, and sets a new standard for regional banking,” chief revenue officer Denise Dresser said in a statement provided to CNBC.

The bank expects to roll out AI agents across lending, deposits and payments over the next six to 12 months.

If they succeed, closing a commercial loan will go from taking 30 to 45 days, including underwriting, document collection and legal negotiations, to about seven days, Sidhu said.

Opening accounts for complex commercial clients, which can take more than a day, will be collapsed to under 20 minutes using conversational AI and automated document gathering, he said.

“When you have an autonomous agent, you’re essentially creating a digital worker … and they can work around the clock,” Sidhu said.

Customers Bank has been laying the groundwork for this announcement for years, first tapping OpenAI in 2023 because Sidhu had what he describes as a tiny investment in the AI giant through his contacts in the venture capital world. The OpenAI deal signed last week broadens their relationship, enabling AI engineers into the bank’s processes, he said.

The bank is among a handful of smaller lenders that target the startup and venture capital community,  and it reportedly bid for Silicon Valley Bank in 2023 amid the regional banking crisis that year.

While it is a relatively tiny firm compared to the likes of JPMorgan Chase, which has $4.9 trillion in assets, Customers Bank has a key advantage, according to Sidhu, who began his career at Goldman Sachs in 2004. The megabanks have sprawling global operations and far higher complexity and regulatory standards for AI implementation, he said.

“Smaller banks are not going to be expected to have the same level of frameworks as many of the larger banks,” he said. Regulators want community and regional banks “to be able to compete with larger banks.”

The lender already uses AI to write half the firm’s software code and has saved 28,000 hours of work so far, equal to not hiring about 15 full-time employees, he said.

“This is an opportunity for us to potentially slow that hiring … and do more revenue per employee,” he said.

The bank is also exploring entering new businesses that would have been prohibitively expensive to tackle before AI agents. For these AI-native business lines, smaller teams oversee automated systems that handle work previously requiring large numbers of humans, he said.

Unlike typical software licensing agreements, Sidhu said both sides are contributing resources to build new tools together, with OpenAI gaining real-world use cases inside a regulated financial institution.

“It’s going to benefit our investors. It’s going to benefit our customers,” Sidhu said. “Our regulators will hopefully also be happier over time, because they’re going to see us reducing risk as well.”