03-21-2026Price:

The Frontier

Your signal. Your price.

AI & TECH

AI agents will replace routine GP visits, says Abridge CEO

Saturday, March 21, 2026 · from 1 podcast
This Week in AI
  • AI models will handle intake and diagnosis for common conditions like rashes and colds, becoming the first point of medical contact.
  • The primary healthcare bottleneck is clerical work: AI agents will manage documentation, coding, and referrals, freeing doctors.
  • New York's ban on LLM medical advice signals regulators recognize this shift is inevitable, not preventable.

Doctors need 30 hours in a day to do their jobs. That's not hyperbole - it's from an American Journal of General Internal Medicine study, and it explains why GP visits consume 20% of healthcare costs.

Abridge CEO Shiv Rao laid out the impossible workflow. Cardiologists prep charts on Sunday nights, then spend consultations typing notes with their backs to patients. They code for billing, fight insurance prior authorizations, and manage referrals - all while trying to deliver care.

Rao sees AI agents coordinating the entire care continuum. One handles intake for routine conditions, another prepares the doctor, a third documents conversations, and a fourth manages post-visit orders. This team tackles what Rao calls "all the jobs" crushing clinicians.

The real obstacle is misaligned incentives. Healthcare operates like pre-Nadella Microsoft - siloed entities pointing guns at each other rather than aligning around patient outcomes. Regulation like New York's LLM ban acknowledges the coming seismic shift, trying to manage what can't be stopped.

When asked whether a family member should see a bottom-tier GP or consult top AI models, Rao's answer was immediate: always the models first, then figure out who to see.

Shiv Rao, This Week in AI:

- I would always do the models and then figure out who to see.

- Two choices: go to the lower third of a general practitioner's and get advice, or get it from the top 3 or 4 models.

ModelsHealth

Source Intelligence

What each podcast actually said

This Week in AI
This Week in AI

How Abridge Built A $5B AI Healthcare Unicorn | Shiv Rao, CEO - This Week in AI Ep 5 • Mar 18

  • Shiv Rao argues that large language models will replace routine medical consultations for common conditions like rashes and colds.
  • A study in the American Journal of General Internal Medicine calculated that doctors would need 30 hours per day to complete all currently required tasks, a workload that Rao says explains why 20% of healthcare costs come from GP visits alone.
  • Rao envisions AI agents coordinating care across the entire continuum, handling patient intake for routine conditions, preparing the doctor, documenting conversations, and managing post-visit orders.
  • The primary obstacle to AI-driven healthcare transformation is not technological but systemic, with misaligned incentives creating a landscape Rao compares to pre-Nadella Microsoft, where siloed entities work against each other instead of aligning around patient outcomes.
  • New York's recent ban on medical advice from LLMs signals, in Rao's view, regulatory recognition that the shift to AI-augmented care is inevitable, not something that can be prevented.
  • When asked to choose between a lower-tier general practitioner and a top AI model for initial medical advice for a family member, Rao stated he would always consult the models first to determine who to see.
  • Current physician workflow, as described by Rao, forces cardiologists to prep charts in their personal time, spend consultations typing notes with their backs to patients, and battle insurance bureaucracy, all while trying to deliver care.
AI & TechHealthRegulation

Related Stories