Latest / Minds, Bodies, and Terawatts / UnitedHealth's $3 Billion AI Push Has Bots Calling Doctors

Transcript

UnitedHealth's $3 Billion AI Push Has Bots Calling Doctors · Minds, Bodies, and Terawatts

• 0:00
  [Host] This is Minds, Bodies, and Terawatts, where the hold music is about to get smarter than the receptionist. Bloomberg reported Thursday that UnitedHealth is investing three billion dollars in AI over the next two years, and it's already testing bots that call your doctor's office to schedule your appointments.
• 0:18
  [Guest] Three billion over 2026 and 2027. They've deployed a chatbot called Avery to six and a half million members so far, expanding to twenty million by year-end. Sandeep Dadlani, their chief digital officer, told STAT News they've, quote, really doubled down since generative AI arrived.
• 0:35
  [Host] So when I call to schedule a checkup, I might be talking to a bot. And when the insurance company calls my doctor, that's also a bot?
• 0:45
  [Guest] Bots on both sides of the phone call. The AI reads medical charts aloud to nurses driving to home visits. It selects billing codes. It flags fraud. UnitedHealth says they're getting a two-to-one return, every dollar in AI saves two in labor costs.
• 1:02
  [Host] Two-to-one on three billion is six billion saved. Where does that money come from?
• 1:07
  [Guest] From the people who used to do those jobs. Claims processors, appointment schedulers, customer service reps. Nobody gets laid off in a press release. The positions get engineered out of existence one function at a time.
• 1:20
  [Host] But this is the same company that faced a class action over an AI called nH Predict that was allegedly denying elderly patients' care with a ninety percent error rate. Now they want us to trust three billion more in AI?
• 1:36
  [Guest] That's the tension. They frame this as taming backlash, faster approvals, fewer errors, less time on hold. But Klarna tried something similar, deployed AI to replace seven hundred customer service agents, then had to rehire humans within a year because the bot couldn't handle nuance.
• 1:53
  [Host] So the easy calls go to the machine and the hard calls go nowhere?
• 1:58
  [Guest] That's the risk. When the machine is both cheaper and reliable enough, managers stop keeping humans as backup. For scheduling, billing, basic claims, UnitedHealth has crossed that line. The open question is whether they push past it into clinical decisions.
• 2:13
  [Host] Does any of this make the actual care better, or just cheaper?
• 2:18
  [Guest] The company says faster. Fewer claims sitting in queues, shorter hold times, less paperwork for doctors. But that's the pitch every corporation makes when it automates the service layer. The real test is whether a patient with a complicated case can still reach a human who has the time and authority to help.
• 2:38
  [Host] What happens to the people whose whole career was being that connective tissue?
• 2:43
  [Guest] Picture a call center floor. One year, every desk has a person and a headset. Next year, half the desks sit empty. Nobody flipped a switch. The floor just got quieter.
• 2:55
  [Host] Patrick Deglon calls this the Labor Cliff in Unscarcity, the point where automation destroys jobs faster than new ones appear. UnitedHealth is pushing that cliff through the largest health system in America.
• 3:07
  [Guest] And that is the story. Not three billion in technology. The quietest workforce reduction in American healthcare, happening one automated phone call at a time. For more on the forces reshaping work, head to unscarcity dot ai. The book is on Amazon, the audiobook is on Spotify. This has been Minds, Bodies, and Terawatts.