How First-Time Founders Can Avoid Mistakes | Pelu Tran, Ferrum Health | EO

Why hospitals need a deployment layer for clinical AI

Ferrum Health is the deployment layer for clinical AI, so hospitals can adopt tools from many vendors without negotiating with each one. Tran arrived at the idea after Augmedix nearly died with Google Glass and after his own family lived through a missed cancer diagnosis. He built for an enduring problem rather than a technology trend, and hospitals are now shopping for exactly the infrastructure he spent five years assembling.

Takeaways

• Start from the pain point, not technology. Tran says the idealistic Stanford graduate in a garage who promises better outcomes fails, and that entering a new technology looking for a problem fails too. Augmedix survived because its technology happened to land on a big problem.
• Ferrum Health deploys clinical AI for hospitals. The company had closed a $16m Series A and raised $31m in total. It consolidates AI vendors into one deployment point, so health systems adopt clinical tools without negotiating with every developer separately.
• A three-day pilot found 50 missed cases. Ferrum ran an open-source lung nodule tool across scans at hospitals in Asia and found almost 50 patients with significant findings their doctors had missed. Tran's uncle died of late-stage lung cancer missed for nearly five years.
• Founders must expect years of being wrong. Keeping the team, investors, stakeholders and customers motivated while nobody cared about the problem was the hardest part. Tran says more patients died from delayed care than from COVID, and Ferrum held its focus for five years.
• Partners describe problems, not solutions. First-time founders ask hospitals what they need, but health systems often cannot solve what they describe. Ferrum tells hospitals upfront that it addresses security and patient safety, and turns away those that lack those problems.
• Hype cycles form around technology, not pain. Tran is careful around new technology trends because boom and bust cycles form around emerging technologies and never around enduring problems. He built Ferrum around a problem that endures rather than chasing the trend.

Augmedix's pivot away from Google Glass

1. Tran started Augmedix as a side project while a medical student at a Bay Area county hospital moving from paper to electronic records. Doctors were burning out, spending three to four times as much of their day on documentation.
2. He and his co-founders built around Google Glass so doctors could stop documenting and spend their day on patients, raising a round from the venture firms on Sandhill Road.
3. Google Glass failed. Because the mission was clear and the point was what the glasses solved for healthcare, Augmedix pivoted to tablet and phone software doctors could carry in a pocket.