Over the past few years, conversational chatbots powered by artificial intelligence are taking their place in health care. You can find dozens of health care chatbots in the market, such as Youper, Ada Health and Florence, among others.

Health care chatbots have countless uses. The top five can be making an appointment, handling simple inquiries, appointment reminders, symptoms interpretation and diagnosis recommendations, while also connecting users with real doctors.

However, health chatbots in the current market are not perfect. Here are three ways to boost the performance of health chatbots and improve customer satisfaction:

Keep training chatbots with new information

One of the reasons why some people prefer human contact to chatbots is because of the artificial “unintelligence.” Sometimes when you put a question in the chat box, the robot responds to you with a few options but none of them is the one you want. After a few rounds, you finally get to speak with an agent.

In those situations, impatient people may feel frustrated. As a result, next time, they would skip the chatbot part and directly say: speak with an agent.

More importantly, in the health care field, if chatbots are not trained well, they can overload the user with possible conditions and cause unnecessary anxiety.

JU law dean: The classic conundrum presented by ChatGPT, especially in academia

Healthful news: What are Jacksonville-area health systems up to?

Letters: City should be more transparent with return on investment for downtown projects

Make chatbots more sensitive to mayday call

During acute illnesses (a stroke, for example), every minute counts. The patient will have sudden confusion or sudden trouble speaking. He or she can only say one word or two or just make some sound.

In these situations, if health chatbots can sense the patient’s condition with the weak signals and immediately call 911, a precious life can be saved. This function is even more crucial for seniors who live alone.

Make chatbots easy to use

Navigating the U.S. health care system might be a challenge for senior citizens, low-income populations and immigrants, due to their health literacy levels, health insurance coverage and language barriers.

Health chatbots are mainly aiming to increase connections between a vulnerable population and health services. However, those populations may also be the ones who know less about technology and are not confident in using an application to address their health problems.

To solve the problem, first, public health programs can be held to advocate the benefits of health chatbots. Second, chatbot companies may need to provide customers with step-to-step easy-to-follow instructions in multiple forms, such as videos, handbooks and voice books, and in multiple languages.

Chatbots have a bright future in health care. By understanding user experience, needs and preferences, chatbots can reach more people, ease the burden for medical providers and promote health equity.

Yining Pan is a graduate student in the University of North Florida’s Master of Public Health program.

This guest column is the opinion of the author and does not necessarily represent the views of the Times-Union. We welcome a diversity of opinions.

This article originally appeared on Florida Times-Union: Chatbots in health care are taking off but still need work