It's Going to Be the Little Things

We are not interested in tokenmaxxing or vibecoding. We don't think an agent replying to emails and organizing to-do lists will make anyone's life meaningfully better. And we don't think AI is going to replace every job, but, like the industrial revolution, it will change how a lot of people earn their living. The largest impact we think AI will have will be on high-stakes American institutions that have become bogged down by inefficient systems and processes.

Consider healthcare, one of several industries where Trase is deploying turnkey agents to replace outdated systems and tools that constrain progress. Duke Health’s Division of Cardiology receives 5,000 faxes a month, each one painstakingly sorted by a nurse or a medical assistant—highly trained and educated individuals who spend a big chunk of their week sifting through papers. This is time that could be spent helping patients.

The fax machine: It’s a relic of bureaucracy that exists for many reasons—interoperability, legacy workflows, sensitive patient data, and a desire for every action to be traceable. All real barriers that burden real staff who face a real limit of 24 hours in a day. Trase exists to fix this problem.

When we started working with Manesh Patel, MD, Chief of the Division of Cardiology at Duke Health, and his staff to figure out where AI agents could lighten the load, the first thing we did was sit with the people actually doing the work, asking them to walk us through a time-consuming process and the systems they touch. How long does each step take? What drives you crazy? What do you want to do more of? Less of? How and where could your valuable time be better spent?

Dr. Patel calls what emerged from those conversations the "liberation layer." Before you can change a system, you have to free the people inside it from work unrelated to their actual jobs. Routing a fax isn't practicing medicine. Chasing a refill for a blood pressure medication someone's been on for six years isn't practicing medicine. These are necessary tasks that don’t merit the time talented professionals spend on them. So that's where we started.

On a Monday morning in mid-April, we turned on Trase’s fax routing agent for the first time to start churning through more than 800 faxes waiting to be processed. By the end of the day, the agent had done almost all of them. In the weeks that followed, inbound fax triage was 7.1x faster, returning 1,395 hours a month to clinical staff. Accuracy went from roughly 70% before the agent to 98.1% after the agent. We’re working with Dr. Patel and his team on peer-reviewed results to document the entire process from baseline collection through impact.

The fax routing agent is just one simple piece of bureaucracy that was tackled meaningfully with Trase. The hard part was building the infrastructure for it to run reliably inside an institution that can't afford to be wrong: data sovereignty requirements, real-time human-to-agent learning loops that continuously improve accuracy, security and audit mechanisms that meet clinical compliance standards, and a judgment layer that determines which decisions are escalated to a person and which aren’t. That work isn't specific to just fax routing. It's the foundation every subsequent agent runs on. We built Trase Origin, our operating system, so that the hardest problems—sovereignty, trust, governance, predictability—get solved once and for all, and every agent that follows inherits them. The result is a platform that manages agents like employees: the same access controls, the same audit trail, and the same compliance requirements as a human staffer. When an agent is uncertain, it escalates. Every decision is logged. No black box.

More agents are in the queue at Duke to address additional administrative workflows as well as other tasks like chronic condition management, including heart failure, hypertension, and high cholesterol. These solutions will not only serve to enhance the clinical experience at Duke by making it meaningfully faster and more efficient; they will help to extend the impact of Dr. Patel and his staff.

We are not doing this alone. Today, Bob Nelson and ARCH Venture Partners joined the team and led our $107M seed round. ARCH has spent nearly four decades backing companies that reshape healthcare and share the belief that complex systems like Duke stand to benefit the most from agentic AI. This funding will allow us to continue to invest in our product and tech and expand our go-to-market efforts.

The fax problem sounds small. In the full scope of American healthcare, it is. But it's where trust gets built—where you prove the infrastructure works before you put it somewhere that matters more. That's our entire philosophy. Start with the thing that drives people crazy and prove an agent can fix it. Then raise the bar, solve a harder problem, and keep pushing the boulder up the hill. The mountain in front of us may seem insurmountable, but with discipline, skill, and the right tools, small advances become large advances that help move our Nation’s mission-critical institutions forward.