a16z Podcast: How Lassie Is Turning AI Into Healthcare Admin Labor

**SPEAKER_1** (0:01)
Welcome to the a16z Podcast. Today we're diving into how power, software, and labor collide in healthcare administration.
Steijn Pelle and Frédéric Renken from Lassie have this fascinating take: AI may be overhyped in Silicon Valley, but it's still underhyped in Iowa. That's such a sharp observation. And Lassie's origin story really proves it. They started with Dr. Quan, a dentist, who was spending two hundred hours a month on paperwork and billing because staff were impossible to find. That pain point turned out to be everywhere. Right, and what struck me is that doctors were nearing burnout not from patient care, but from administration. The founders realized their real product isn't a dashboard—it's labor itself. Exactly. That's why they built Lassie as an agent, not just a tool. They're trying to actually do the work rather than merely store the data. The goal is high autonomy, with only a sliver left for humans. The team even spent months in dental offices acting as the humans in the loop before the system could handle things independently. That hands-on approach is crucial. They stress that good software in this world has to be truly hands-off. It needs to scale, integrate with existing practice systems, and run core workflows reliably.
Their background at places like Robinhood and Superhuman clearly shaped their bar for speed and polish.
But here's where it gets tricky—distribution is a huge theme. This isn't an app-store download problem. It's a last-mile adoption problem for busy, non-technical owners who aren't sitting on LinkedIn all day. How are they solving that? They're building toward fully self-serve setup with a checkout-like flow that connects bank accounts, records systems, and insurance portals. They say they're months away from that vision. The strategic piece is interesting too. Steijn argues that startups should own the pipe, not just the feature, because incumbents can copy visible features quickly. He even references the Tivo problem, saying "It's so crowded, nobody goes there anymore" when the customer relationship sits elsewhere. And that risk is sharper in the AI era, because bad engineering can look much better with AI tools. But they see a big opening in places without a dominant incumbent. In small businesses, the real competitor is often Betty, a retired assistant, or an overseas billing agency. There's also the challenge of teaching models the actual workflow.
Models don't automatically know how to file insurance claims or follow practice-specific procedures, so they have to encode real operations and standard operating rules.
Which makes the product defensible, but deeply manual to build. Still, the bigger vision is bold—every small business running itself, with agents handling busy work so people can focus on their craft.
For dentists, the payoff isn't just efficiency. It's capacity, access, and fewer people leaving the profession because of paperwork.
The future isn't less work for patients to receive care—it's less waste for the people delivering it.