AI coaches get diabetes care started sooner, while people rate human coaches higher
☕ Includes a question for the family ↓
Tools for living with diabetesPart of a story we are following · 4 articles · See every update
If you have prediabetes, waiting to start a program is a common hurdle. A new trial suggests an AI coach can get you going sooner, while a human coach tends to leave people happier with the experience. Both paths work as a way into diabetes care, so you can pick the one that fits you.
In a randomized trial of 368 adults with prediabetes and overweight or obesity, people assigned to a fully automated AI diabetes prevention program started it in a median of 11 days, compared with 26 days for people assigned to a human coach.
How the study worked
Researchers randomly assigned 368 adults with prediabetes and overweight or obesity to one of two versions of a diabetes prevention program. One was run entirely by AI. The other was led by a human coach and recognized by the CDC. They followed engagement over 12 months and asked people to rate their experience and say what they would have preferred. The results appeared in NPJ Digit Med on August 3, 2026, and this report focuses on engagement, ratings and preferences.
What they found
The AI program won on speed. People in the AI group began the program sooner, with a median of 11 days versus 26 days for the human-coach group. “Median” means the middle person: half started faster than that, and half took longer.
The AI group also stayed involved in a steadier way. Over the 12 months, their engagement was higher and more evenly spread out, rather than bunching up in a few busy stretches.
The human program won on how it felt. People rated the human-led program higher in every area the researchers asked about. The biggest gap was in satisfaction.
Preferences pointed the same way. Among people assigned to the AI program, 46.5% said they would have preferred a human coach. Among people assigned to the human-led program, 31.5% said they would have preferred a fully automated one.
So it is not a case of one program beating the other. The researchers describe a trade-off between reaching more people quickly and how good the experience feels to the person going through it.
Why does the speed of starting matter?
A program only helps if you actually begin it. Many people put off signing up because of scheduling, cost or simply not knowing where to start. An AI option that is ready right away can remove some of that waiting. And the steadier engagement in the AI group shows it can keep people involved over the full year.
At the same time, many people value a real person who knows their name and can talk things through. If that is you, your instinct has some support in these results. Nearly half of the people given the AI program wished they had a human coach instead.
What you can do today
Think about which style would keep you going: a person you meet with, an app that is there any time, or a mix. Then look for a CDC-recognized diabetes prevention program near you or online, and ask whether it is led by a human coach, by AI, or both. Knowing your preference before you call makes the first step quicker.
Talk about it at home
In this study, researchers let chance decide who got which kind of program, so the two groups started out much alike. Then they watched how quickly people began and how they felt about it. Think of two ways into the same park: one gate is a turnstile that opens the moment you arrive, and the other is a friendly guide who greets you by name but is sometimes a little late.
Question for the table
Imagine a helper that is ready the very second you ask, and another helper who knows your name but is busy sometimes. Can your family find one good reason to pick the quick helper and one good reason to pick the one who knows you?
Keep exploring
Copies this article and its questions so you can paste them into ChatGPT, Claude, Gemini or the AI you use, and keep the conversation going.
Questions and answers
- Did the AI program or the human coach program work better?
- It depends on what you measure. The AI program got people started sooner and kept engagement steadier. The human-led program earned higher satisfaction ratings in every area asked about.
- Who took part in the study?
- A total of 368 adults with prediabetes and overweight or obesity were randomly assigned to an AI-led or a human-led diabetes prevention program.
- How many people would have preferred a human coach?
- In the AI group, 46.5% said they would have preferred a human coach. In the human-led group, 31.5% said they would have preferred a fully automated program.
- How long did the study follow people?
- Engagement was tracked over 12 months.
