🎙️ Voice is AI-generated. Inconsistencies may occur.

3 min read

Zocdoc Launches AI Phone Assistant

Meet Zo: the AI agent who vows not to put you on a brief hold.

  • Español
  • 中国人
  • Français
  • Deutsch
  • Portuguese
  • हिन्दी

Zocdoc—the digital platform that helps patients find doctors and book medical appointments—is unveiling an AI assistant today to make that process even easier.

The AI agent named "Zo" can handle unlimited scheduling calls from patients, and is trained to speak naturally and conversationally as a human call center employee would.

Only about half of patients are satisfied with the service at their health care provider's call center, according to a 2023 survey of 200 senior leaders. The average hold time at these organizations was 4.4 minutes, well above the HFMA's recommendation of 50 seconds.

Zo, however, can answer calls automatically, eliminating hold times entirely. Zocdoc believes that the new agent can improve the booking experience for both patients and health care organizations, CEO Dr. Oliver Kharraz told Newsweek.

Zocdoc launches AI phone assistant
Zocdoc's new agentic AI assistant, "Zo," can answer phone calls and schedule patient appointments without human intervention. Newsweek illustration/ Getty Images

"My hope is that this is the beginning of a much more pleasing health care experience where we free up call center capacity to help folks with more difficult, off-the-beaten-path requests, and we leave all the rote things to automation," he said.

The new AI agent integrates with medical practices' existing phone systems and electronic health records. There are no upfront fees to try it out, but practices will be charged for Zo's "outcomes," Kharraz said. Each time the AI agent successfully autonomously schedules an appointment, the practice will pay a $2 standard charge.

"The reason why we can do that is because we've been around for nearly 20 years, we have a growing, profitable company, and we know this space exceedingly well," Kharraz said. "We can bear this risk and put our money where our mouth is."

That same depth and breadth of experience helped Zocdoc train Zo. Its website had more than two decades of data that could be leveraged to explain integrations, insurance plans and patient/provider preferences. About 150,000 doctors across all 50 states are customers of the platform.

In addition to vanishing hold times, Zo never has a bad day or a fight with their spouse, Kharraz said. The agent has an unlimited capacity for patience and politeness, which could make it more pleasant to talk to. It could also ease the staffing burden on call center managers—Zo can work alone and around the clock and can easily adjust to spikes and lulls in call volumes.

While many health systems are experimenting with AI, there is still a sense of hesitancy around full-fledged adoption.

In a recent survey of 300 health care executives, only 3 percent of respondents pointed to call centers and customer service as key areas for generative AI transformation. Agentic AI, which can complete autonomous tasks from start to finish, may be a more plausible solution for bottlenecks in these departments.

"We started [Zo] two years ago and we're just launching it now," Kharraz said. "Why the delay? We believe that it is most promising when an AI can autonomously complete a task—that's when you really get the productivity increase out of it."

"It just took us so long to create an agent that we think can do that reliably in the vast majority of cases," he continued. "It's so easy to create a good sounding demo of [agentic AI], because these voices sound great, and whatever it is they say sounds somewhat reasonable. But that's so far from having something that actually works in practice."

Zo struggled with social norms along the way, according to Kharraz. For example, Zo might ask a patient for their insurance card, and the patient may say to themselves, "Where did I put my insurance card?"

In an attempt to be helpful, the earlier iteration of Zo would suggest that the patient check their wallet or ask their spouse. But a human call center employee would typically tell the patient to take their time.

"There were literally thousands of these learnings that we had to identify so that Zo truly behaves the way a human being would expect to interact on the phone," Kharraz said. "I think that's the big difference between Zo and a lot of great sounding demos—it actually has all that confidence built into it."

The Latest in AI

EXCLUSIVE
05.29.25
HEALTH
05.01.25
MOBILITY
04.29.25
MOBILITY
04.22.25
HEALTH
04.23.25
See all Stories