Customers Bank embeds OpenAI engineers to automate lending, onboarding

Customers Bank opened its latest earnings message with a voice that was not Sam Sidhu’s. The bank’s chief executive said the prepared remarks on the April 24 call were delivered by his AI clone, a stunt Sidhu said may have been the first of its kind on a public-company earnings call. It was more than a showpiece. It signaled how aggressively the West Reading, Pennsylvania-based lender is pushing generative AI into the center of how it runs.

The bigger move is a multiyear partnership with OpenAI that will embed OpenAI engineers inside Customers Bank to help automate lending, client onboarding, deposits and payments. OpenAI Chief Revenue Officer Denise Dresser said the company 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.” For a regulated bank, that is a more consequential experiment than a customer-service chatbot. It puts AI closer to credit decisions, account opening and back-office workflow, where controls, audit trails and data privacy matter as much as speed.

Customers Bank said it wants to cut loan processing time from weeks to days and push its efficiency ratio from about 49% into the low 40s. The first-quarter 2026 efficiency ratio came in at 49.68%, down from 52.94% a year earlier, while net income available to common shareholders rose to $69.7 million, or $1.97 per diluted share. Total deposits reached $21.6 billion and total loans were $17.4 billion. Sidhu said the payoff should start showing up in higher returns in 2027.

The bank is also trying to turn AI from a pilot project into a workforce discipline. A recent earnings transcript summary said 75% of employees had AI licenses, staff had built more than 500 agents and custom GPTs, and more than 28,000 hours had been saved, roughly equal to almost 15 full-time jobs. Customers Bank said the aim is an end-to-end, agentic operating model, one that could eventually be strong enough to sell to other banks as a product.

That ambition carries the same risks that now shadow most serious AI deployments in finance. If lending and onboarding are increasingly automated, bank leaders must still answer who owns the model, who reviews exceptions, how customer data is protected, and how a regulator or auditor can trace a decision after the fact. Customers Bank, founded in 2009 after Jay Sidhu and investors recapitalized a small Pennsylvania bank, has spent years scaling fast. With Jay Sidhu retired effective January 1, 2026 and Sam Sidhu now in charge, the bank is betting that AI can widen margins without widening headcount at the same pace.