From single-point solution to multi-agent OS.

Sully.ai began with a clear job: help physicians produce clinical notes. The next opportunity was larger. Could the same company help a healthcare organization manage the work around the visit, not just document what happened inside it?

That meant moving from one useful AI product to a coordinated system for patient access, clinical documentation, decision support, and coding. The challenge was not adding more agents. It was deciding what they should own, what they should share, and how people would remain in control.

As Head of Product & Design, Matthew Belferman led the product vision and capability definition for that transition. He defined the agents’ roles, the work they needed to accomplish, and how people would use and supervise them. Working with Sully’s founders, AI research, engineering, and clinical teams, he connected those decisions to interaction design, technical tradeoffs, and an offering enterprise customers could understand and adopt.

A note is one part of that work. Before it, someone answers a call, identifies a patient, finds an appointment, and prepares the chart. After it, someone checks the documentation, prepares codes, and follows up. Automating one step can leave the next bottleneck untouched.