Customers Bank and OpenAI team up on AI deposit onboarding platform

Customers Bank and OpenAI team up on AI deposit onboarding platform

The Pennsylvania lender adds Palantir to an AI push that already includes a multiyear OpenAI collaboration aimed at faster account openings

Customers Bank is bringing Palantir into its AI plans. The two companies will build an AI-powered platform for deposit onboarding, the process of getting new customers’ money through the door and into accounts.

What the bank is building, and with whom

The Palantir partnership centers on a single job: deposit onboarding. The deal sits alongside a larger AI project the bank has already made public. On April 27, 2026, Customers Bank announced a multiyear collaboration with OpenAI to overhaul its commercial banking operations. Streamlined digital onboarding for deposits was a core focus of that effort too.

The OpenAI arrangement sets out unusually specific goals. The bank wants commercial deposit account openings to take less than 20 minutes, down from more than a day. The goals reach beyond deposits. Customers Bank is aiming to shrink commercial loan closings from 30-45 days to approximately one week. It also wants its efficiency ratio to fall from around 49% to the low 40s, with returns expected to begin in 2027.

Advertisement

A bank that has been at this for a while

Customers Bank is a subsidiary of Customers Bancorp and is based in West Reading, Pennsylvania. It holds approximately $26 billion in assets.

Its AI habit goes back to 2023, when it adopted ChatGPT Enterprise. By the time of the April announcement, 75% of bank staff were using OpenAI-powered tools. Employees have built more than 500 custom agents and models since the relationship began. Those tools have saved 28,000 hours of work, which the bank equates to roughly 15 full-time employees.

The OpenAI collaboration also changes how the work gets done. OpenAI’s technical teams are set to be embedded directly in the bank’s operations. The aim is custom solutions built on the bank’s own data and workflows, rather than off-the-shelf tools.

CEO Sam Sidhu has framed the ambition plainly. He wants Customers Bank to become an “AI-native bank.”

What this means for banks and investors

For Customers Bancorp shareholders, the numbers to watch are the ones the bank has put on the record. Account opening times, loan closing speed and, above all, the efficiency ratio. If that ratio moves toward the low 40s as hoped, AI stops being a talking point and starts showing up in earnings.

There is a potential second act. Building tools alongside OpenAI engineers could produce products the bank might eventually market to other financial institutions, though it remains a possibility rather than a plan with a launch date.

The bank’s stated targets give outsiders a clear scorecard. Under 20 minutes for a commercial deposit account. About a week to close a commercial loan. An efficiency ratio in the low 40s, with returns expected from 2027.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.
Customers Bank and OpenAI team up on AI deposit onboarding platform
Customers Bank and OpenAI team up on AI deposit onboarding platform

The Pennsylvania lender adds Palantir to an AI push that already includes a multiyear OpenAI collaboration aimed at faster account openings

Customers Bank is bringing Palantir into its AI plans. The two companies will build an AI-powered platform for deposit onboarding, the process of getting new customers’ money through the door and into accounts.

What the bank is building, and with whom

The Palantir partnership centers on a single job: deposit onboarding. The deal sits alongside a larger AI project the bank has already made public. On April 27, 2026, Customers Bank announced a multiyear collaboration with OpenAI to overhaul its commercial banking operations. Streamlined digital onboarding for deposits was a core focus of that effort too.

The OpenAI arrangement sets out unusually specific goals. The bank wants commercial deposit account openings to take less than 20 minutes, down from more than a day. The goals reach beyond deposits. Customers Bank is aiming to shrink commercial loan closings from 30-45 days to approximately one week. It also wants its efficiency ratio to fall from around 49% to the low 40s, with returns expected to begin in 2027.

Advertisement

A bank that has been at this for a while

Customers Bank is a subsidiary of Customers Bancorp and is based in West Reading, Pennsylvania. It holds approximately $26 billion in assets.

Its AI habit goes back to 2023, when it adopted ChatGPT Enterprise. By the time of the April announcement, 75% of bank staff were using OpenAI-powered tools. Employees have built more than 500 custom agents and models since the relationship began. Those tools have saved 28,000 hours of work, which the bank equates to roughly 15 full-time employees.

The OpenAI collaboration also changes how the work gets done. OpenAI’s technical teams are set to be embedded directly in the bank’s operations. The aim is custom solutions built on the bank’s own data and workflows, rather than off-the-shelf tools.

CEO Sam Sidhu has framed the ambition plainly. He wants Customers Bank to become an “AI-native bank.”

What this means for banks and investors

For Customers Bancorp shareholders, the numbers to watch are the ones the bank has put on the record. Account opening times, loan closing speed and, above all, the efficiency ratio. If that ratio moves toward the low 40s as hoped, AI stops being a talking point and starts showing up in earnings.

There is a potential second act. Building tools alongside OpenAI engineers could produce products the bank might eventually market to other financial institutions, though it remains a possibility rather than a plan with a launch date.

The bank’s stated targets give outsiders a clear scorecard. Under 20 minutes for a commercial deposit account. About a week to close a commercial loan. An efficiency ratio in the low 40s, with returns expected from 2027.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.