In 2025, J.P. Morgan Payments deployed a first-of-its-kind mail automation robot at one of its lockbox sites, marking an inflection point in the decades-long evolution to modernize check processing.

Many companies and organizations use a bank “lockbox” service to streamline incoming payments. Instead of mailing checks to an organization’s office, payers send checks and accompanying remittance documents (such as invoices or payment stubs) to a secure address managed by the bank. The bank then receives the mail, deposits the checks, and digitizes the accompanying documents so the company can quickly reconcile payments.

Developed with Ripcord, the Silicon Valley-based document intelligence company, the solution streamlines these lockbox operations. Powered by AI models, these highly specialized robots can open envelopes, remove and unfold the contents, separate pages, remove staples, track content and scan the underlying documents, automating work lockbox operators previously handled manually across roughly 480 million documents and checks each year.

Engineered for real-world complexity, the robots – powered by computer vision and AI models – can handle more than 4,000 envelopes and document permutations, including differing layouts and content types. This level of variability is trivial for humans but extraordinarily challenging to encode for machines.

“Lockbox remains a cornerstone of our client offerings,” said Michelle Conklin, Head of Receivables and Public Sector, J.P. Morgan Payments. “By investing in robotic and AI technology to improve our lockbox operations, we are automating the most labor-intensive tasks of the process, freeing our team to focus on more complex, higher-value decision making. The result is a faster, more secure, and smarter receivables process that gives our clients both agility and peace of mind.”

This isn’t the first time we’ve applied AI solutions to modernize our lockbox operations. In 2020, we undertook a significant step by rebuilding the underlying processing platform. The internally developed solution uses artificial intelligence across core processing tasks, including data extraction, quality validation, and review.

To complement these solutions, we have integrated large language models to automate complex review tasks with greater speed and precision. We are using automation to take on the most human-intensive parts of lockbox processing – without trading off quality. Operators now have better tools and real-time insight into throughput, exceptions, and turnaround times – visibility that didn’t exist before – so teams can keep improving the process over time.

The results speak for themselves:

  • Data capture that previously required approximately 13 billion keystrokes for lockbox operators each year is now largely automated.
  • Processing accuracy has increased to 99.999%+, meaning more precise information delivered to clients.

Lockbox employees now have an AI-powered assistant to obtain real-time process-related information, so operators can stay focused on the work that requires human decision-making and expertise.

As a global and highly regulated entity, compliance and security are also built into the development processes from the start. Every model undergoes extensive testing before deployment.  

This approach lets us layer innovations on top of one another, so each improvement builds on the last. The result is compounding processing power and more opportunities for us to support additional clients, use cases, and industries. It’s backed by our broader $19.8 billion annual technology investment, a firmwide commitment that funds technology across the company, including initiatives like this one.

Visit J.P. Morgan Receivables to learn more.

© 2026 JPMorgan Chase & Co. All rights reserved. JPMorgan Chase Bank, N.A. Member FDIC. Deposits held in non-U.S. branches are not FDIC insured.  Non-deposit products are not FDIC insured. The statements herein are confidential and proprietary and not intended to be legally binding. Not all products and services are available in all geographical areas. Visit jpmorgan.com/paymentsdisclosure for further disclosures and disclaimers related to this content.

References

1.

J.P. Morgan Internal Data, 2025