How does healthcare use virtual agents and chatbots?

A new tool to tackle healthcare issues is stepping up to the plate: virtual agents.

Virtual agents, also known as chatbots or AI voice agents, are offered by electronic health record (EHR) giants including Epic, Athenahealth, and Oracle as well as major US insurers like UnitedHealthcare and Elevance to help users with tasks such as estimating healthcare costs and scheduling appointments.

These agents communicate with patients via voice or chat in order to lighten administrative load during staffing shortages and save money amid financial strain. Some healthcare leaders even see the potential for AI voice agents to help improve access to care and patient outcomes. In a Q4 2024 survey from consultancy McKinsey, 55% of healthcare leaders said one of generative AI’s greatest potentials was in improving patient and member engagement.

“We see these technologies as a catalyst for transforming care,” Sonja O’Malley, the general manager of business development and licensing at Cleveland Clinic Innovations, told us.

The worldwide healthcare AI voice agent market was estimated at $468 million in 2024 and is predicted to reach over $3 billion by 2030, per Grand View Research.

How it works

When you think of a virtual AI agent or assistant, you might mistakenly think of their predecessor: interactive voice response (IVR) systems. IVRs, or phone trees, are scripted programs with preset responses that are triggered by specific words or actions, like pressing a number on a phone’s keypad, and trace their origins back to the 1930s.

In contrast, AI agents can have wide-ranging conversations with patients and even communicate with other AI agents to accomplish tasks. They are typically grounded in large language learning models and defined by their ability to use tools, act autonomously, self-reflect, and adapt, John Licato, associate professor of computer science and engineering at the University of South Florida, told us.

“[AI] becomes agentic when we give it the ability to decide what to do beyond just a fixed set of canned responses,” Licato said.

Agentic AI also has the ability to remember details from phone calls that happened months before, Ankit Jain, CEO of agentic healthcare communications platform Infinitus Systems, told us.

“Even your best healthcare worker, no matter how amazing they are, they’re overworked. There’s no way they can keep track of every little detail for every member out there,” Jain said.

How it’s used

One of payers’ and providers’ main uses for AI voice agents and chatbots is in call centers.

AI agents can help automate things that previously required a person, like scheduling appointments or answering basic questions. This helps reduce annual call center costs—which ran an average of $13.9 million, per a 2023 industry report by conversational AI platform Hyro—and get people scheduled sooner for appointments, Jessica Lamb, a partner at McKinsey, told us.

“Call center is a huge area of spend,” Lamb said. “It is a great application of a voice agent.”

Tampa General Hospital saved an estimated $67,000 per month by automating the ability to cancel, reschedule, and confirm appointments in December, Amy Bender, the hospital’s director of IT digital experience, told Healthcare Brew. UnitedHealthcare CEO Tim Noel similarly touted AI as a tool to create “higher customer experience and satisfaction at a lower cost” during a Jan. 27 earnings call. He said 80+% of calls use AI tools to answer members’ questions “faster and more accurately.”

Getting ambitious

Health systems are also taking AI’s capabilities beyond just scheduling appointments and answering questions by using AI agents in preventative care initiatives. Cleveland Clinic, for instance, uses AI voice agents and chatbots to check in with patients who have chronic conditions like hypertension, O’Malley said.

“We are able to reach out to the patients in advance of any escalating situations, understand how they’re doing, understand if they are getting the care that they need,” she said.

Cleveland Clinic also uses AI agents to help with a preventative screening program for patients with a health history that suggests they are at risk of developing lung cancer. AI agents call these patients, tell them about the program, and help them navigate screenings and follow-up care, O’Malley explained, adding the health system has seen an increase in patient engagement in the program compared to only using traditional methods like letters and messages on a patient’s EHR portal, she added.

“We can not only democratize care but [also] reach more patients and care for them proactively, and even in a predictive manner,” she said.