AI-Native Healthcare: 100M Doctor Visits, 10–20 Hours Saved, Prior Auth in Minutes — Janie Lee & Chai Asawa, Abridge

In this episode of Latent Space: The AI Engineer Podcast, hosts dive deep into Abridge, a clinical intelligence layer transforming healthcare through ambient AI. Co-founders Janie Lee and Chai Asawa discuss how their platform began by reducing clinician documentation burden—saving doctors 10–20 hours a week—before evolving into a broader intelligence layer that improves care delivery, reduces costs, and saves lives. They highlight the shift from reactive alerts to proactive, context-aware interventions, such as real-time prior authorization guidance that can approve medical imaging before a patient leaves the room. The conversation explores the technical and operational challenges of building AI in healthcare: managing massive data complexity across EHRs, payer policies, and specialties; ensuring model accuracy and low latency; and maintaining HIPAA compliance through de-identification and secure data handling. Abridge’s unique moat comes from its 100 million+ real-world clinical conversations, which fuel a data flywheel for continuous improvement. The team emphasizes the importance of clinician scientists, rigorous evaluation frameworks, and progressive rollout to build trust with health systems. They reflect on how healthcare’s high stakes are driving AI innovation faster than other domains, and how their product is uniquely positioned to unify providers, payers, and pharma through a shared clinical intelligence layer. Key takeaways include: 1) AI in healthcare must prioritize context, safety, and real-time performance over flashy prototypes; 2) The doctor-patient conversation is the central data source and workflow—everything else is derivative; 3) Success comes from deep integration with EHRs, personalized workflows by specialty and organization, and a focus on reducing latency in care; 4) Trust is earned through incremental rollout, transparency, and operational rigor; 5) The future lies in agentic systems that act on behalf of clinicians, not just alert them. The episode ends with a vision of Abridge as a foundational layer for a more efficient, equitable, and intelligent healthcare system.