Customers Bank Engages OpenAI to Reengineer Commercial Banking

Customers Bank announced a multiyear strategic collaboration with OpenAI to deploy artificial intelligence across its commercial banking operations, according to a BusinessWire press release published April 27, 2026. The release states this expands a relationship that began in 2023 when Customers Bank adopted ChatGPT Enterprise. Customers Bancorp, the bank's parent, is described in the release as having nearly $26 billion in assets. The announcement identifies three initial focus areas, lending, deposits and payments, and says the engagement includes direct onsite work with OpenAI technical teams to build custom AI capabilities on Customers Bank's processes, data and institutional knowledge. The release also states that 75% of the bank's employees use tools powered by OpenAI.

Reporting across the press release, PYMNTS, and Pottsville Mercury describes use cases the bank intends to prioritize. In lending, the bank plans to apply AI for document collection, credit memoranda preparation, legal documents, and portfolio/collateral monitoring, per the announcement. For deposits, the release highlights streamlining digital onboarding and account setup. For payments, the announcement says AI will accelerate the capabilities of the bank's proprietary payments platform, cubiX. Quartz additionally reports the bank aims to reduce commercial loan closing times from 30-45 days to about 7 days, which would require automated data ingestion and workflow orchestration to be effective.

Editorial analysis: Financial firms moving beyond off-the-shelf chat tools toward embedded, custom model integrations are increasingly common. Companies that combine vendor model access with on-premises or tenant-isolated deployments tend to emphasize data governance, access controls and risk management, which Customers Bank highlights in its release. For enterprise ML engineers, this pattern raises integration, observability and model governance requirements across core banking systems.

Editorial analysis: Implementing the announced use cases typically requires robust document ingestion pipelines, production-grade OCR, structured extraction, entity resolution, privacy-preserving fine-tuning or retrieval-augmented generation, and end-to-end audit logging. Firms attempting similar projects usually allocate effort to latency, throughput and explainability for credit decisions to meet regulatory scrutiny.

Editorial analysis: Observers should track:

• •whether Customers Bank publishes technical guardrails or model validation metrics
• •concrete time-to-close improvements or pilot results for commercial loans
• •how the bank maps AI outputs into regulatory reporting and audit trails. Public progress on those indicators will show whether the collaboration yields measurable operational gains beyond vendor partnership announcements