A personal injury case management company faced a tangle of email, a home-grown case management system, and complex bookkeeping. Their staff spent entire days manually entering healthcare provider invoices into multiple systems—a "swivel-chair" process that limited the company's capacity to manage more cases.

The company's CPTO had worked with Kelsus before and trusted their ability to move fast, understand business processes, and deliver startup-level engineering speed with enterprise-grade reliability.

Kelsus built an invoice automation system that:

• Pulled invoices directly from email (individually or in bulk).
• Parsed them using AI (image-reading LLMs and AWS Textract).
• Automatically created entries in both the case management and Microsoft bookkeeping systems.
• Flagged difficult invoices in an exception queue for quick human review.

One batch of 80,000 invoices was processed in under an hour, a task that would have taken the bookkeeping team more than two weeks.

Staff members immediately felt the impact. As one client note put it: "The automated invoicing stuff has dramatically improved [our team member's] quality of life… it greatly improved speed, accuracy, and highlighting errors. I'm pumped about what we can automate next!"

Building on this success, Kelsus is now working with the client on machine learning models to support underwriting decisions—helping them predict which personal injury cases are most likely to follow through to settlement.