Conversations about our mental health and the impact of technology on our wellbeing have been heating up in recent years, but there’s another field that seems to be quietly weighing in on how technology can improve quality of life: artificial intelligence.

Over the past year, researchers and academics at IBM Watson, Microsoft, and Google have worked to apply A.I. to the quality of communication in health care, with early results including benefit for both patients and healthcare organizations.

Using A.I. to exchange positive feedback between patients and health care providers is one of the few areas in which A.I. is already used in a tried-and-true way, in commercial products such as voice assistants such as Alexa and Google Home. However, given the vast amount of ongoing data that could be gathered—such as the voice volume, keywords, and the context of how each message is being delivered—it may be one of the largest applications yet. The prevalence of smartphones, in particular, and text messages, in particular, may have given this method of outreach an even greater chance at success.

The Case for Realtime Innovation

Humans are innately adept at keeping score: the score of a tennis match, the speed of an old woman’s walk, and the quality of honey. As technology gets better at using data to personalize the experience, we humans are increasingly trying to learn from what data tells us about human behavior. Yet understanding the type of communications a human receives from her loved ones or a healthcare professional can help us identify the types of interactions that are not only meaningful, but are important in shaping a person’s outlook on their health. We hear words such as “social feedback” and “rewarding therapy” frequently in relation to health care.

At IBM Watson Health, researchers at IBM’s research labs and inside IBM Watson Health and IBM’s Watson Group spend a lot of time exploring ways to create real-time feedback for organizations. We build advanced A.I. solutions that seek to provide positive social feedback to the masses and truly transform the healthcare experience. As part of our work, we have been analyzing millions of text messages on texts that occurred between people with similar health issues as others with similar health problems. We have also conducted research using GreetSOS, a voice messaging service that allows people with chronic diseases to quickly message their health care providers, in order to quantify the positive impact on patients. While this type of research on text has been ongoing for years, what has been working today for us has a strong learning component.

One of the most interesting findings in our most recent research was in how text messages can help in the conversion of positive emotional messages to positive behavior. As individuals communicate with their health care providers, some of them may seek to share their positive mental health experiences, or feel emotional distress. When they do this, their messages will usually be delivered with a very general approach, such as, “I feel very good today. I was feeling like crap yesterday.”

While this approach may be fine for a conversation with a friend or a family member, it is far less effective when communicating with a professional. However, it is important to remember that we are trying to generate information, and we always want to get to the end goal of getting a patient to commit to improving their health. Since we want people to follow through on their positive recommendations, we have to develop strategies for individuals that fit them.