Changing behaviour is hard—even when people genuinely want to. Whether it’s exercising more, eating healthier, or making other lifestyle changes, sticking to new habits is often more challenging than expected.
“Adherence”—sticking to behaviour change—is arguably the most important unsolved puzzle affecting our health and well-being. Poor adherence—whether in preventive care or problem-based treatment - costs $300bn annually in the U.S. alone. Even in health tech, where apps promise to make improving well-being easier, the average user drops off after just four days.*
This problem isn’t limited to healthcare; poor adherence affects businesses across industries, leading to disengagement and lost revenue. While increasing adherence is a key goal of many companies, it remains elusive. So, how do we fix it?
One strategy companies have used to solve the adherence problem is adding a digital coach as a core feature in their apps, but this has had limited success. Why? Digital coaches have tended to use an algorithmic, one-size-fits-all approach without tapping into the power of behavioural science research.
When creating Sproutiful, our AI proof of concept, we took a different approach. By combining AI’s ability to learn and adapt with the proven effectiveness of science-backed coaching, we aimed to create a digital coach that offers connection, customisation, and genuine insight. As a behavioural scientist, doctor, and coach, I joined the team to help ensure the app used the best behavioural science at every step.
The science-based technique at the core of Sproutiful is Motivational Interviewing (MI)—a proven, evidence-based coaching technique that taps into intrinsic motivation. MI uses empathic, open-ended questioning to help people clarify their goals, resolve internal conflict, and find their own reasons for change. With over 1,200 studies backing its effectiveness, MI has long been used in the clinical analogue world, but was previously thought too complex to scale in an app without direct human involvement.
AI has now made this possible. By continuously learning from users and offering personalised, empathetic guidance, Sproutiful moves beyond the transactional nature of traditional digital health tools. This blend of AI and MI creates a dynamic and genuine relationship with the user. This new kind of relationship with users could revolutionise how digital health delivers long-term behaviour change, and ustwo is leading the charge.
Applying the science of Intrinsic Motivation
To make the Sproutiful coach truly science-backed, it was important to lean on principles with the strongest evidence behind them. Research overwhelmingly demonstrates that leveraging intrinsic motivation—the drive to do something because it’s enjoyable, feels right and is in line with our own values, identity and personal goals—is the most effective method to encourage lasting positive behaviour change.
In contrast, depending primarily on extrinsic motivation, which focuses on factors like others’ opinions and expectations of us, is often counterproductive. A doctor, app or book telling us we should do a certain thing can be effective in the short term, but is not a recipe for long-lasting change.
Motivational Interviewing, a widely researched technique, leverages intrinsic motivation at its core. The questions it asks first and foremost help a person find their way back to what they care about, not what others care about for them. This approach has been shown to increase a sense of autonomy, self-efficacy and overall motivation to stick to a behaviour change.
While research studies can make Motivational Interviewing seem technical, when you experience it in-person, it is deeply human. I’ve seen its powerful impact up close for more than 12 years, using it with everyone from patients to executives.
An AI Coach that asks the right questions
Good coaching always starts with asking good questions that meet the user where they are. Today, AI like ChatGPT has focused primarily on answering questions rather than asking questions. The Sproutiful coach flips this model on its head. Instead of barking orders or giving top-down advice, it uses a range of MI techniques to ask open-ended questions, validate the user’s experience, and uncover genuine insight.
Inquiry, not advocacy
Open-ended questions empower users to feel ownership over their journey. Rather than dictating actions like “Eat more plants for gut health,” Sproutiful asks, “How are you feeling about eating more fruits and vegetables?” Even closed-end questions put the user in the driver’s seat, for example, “On a scale from 1 to 10, how motivated are you to improve your gut health?”
Mirroring that ‘rolls with resistance’
By mirroring, or reflecting back what users express in new ways, the coach helps them process feelings and gain insight. For example, Sproutiful might say, “I hear you don’t want to make changes that feel too big right now because there’s a lot on your plate,” validating the user’s thoughts. Instead of pushing back, the coach “rolls with resistance”—accepting the user’s hesitation without arguing, which research shows often makes people more open to change.
Increasing ‘change talk’
Sproutiful promotes “change talk,” a method of encouraging users to express their own reasons for change, making them more likely to stick to new behaviours. Rather than simply saying, “Eat more plants for your gut health,” the coach asks questions like, “What benefits from increasing your veggie intake would excite you the most?” or “Can you share a time when you enjoyed eating vegetables?”
Supporting self-efficacy
Self-efficacy—the belief that one can succeed—is crucial for sustaining behaviour change. Sproutiful supports this by asking users about their past successes and encouraging small, manageable steps. If a user struggles with eating more vegetables, the coach might say, “Can you recall a previous time when you wanted to make a difficult change and you found a way to stick with it? Is there anything you learned then that feels relevant now?”
Resolving internal conflict
Sproutiful helps users resolve internal conflicts that often hinder behaviour change, like wanting to eat healthier while craving comfort foods. By asking reflective questions such as, “What desires get in the way of achieving your goals?”, the coach helps users clarify their priorities and reduce ambivalence. It also highlights discrepancies between users’ actions and goals, gently pointing out gaps without judgement, for example, “If 10 out of 10 is where you want to be with respect to your plant intake, where are you on that scale now?” Such questions increase clarity and resolve.
At every step, Sproutiful integrates these science-based principles without overwhelming the user with technical jargon or rigid techniques. Instead, it offers a UX that is playful, human-like, and approachable.
The surprising overlap between UX design and science-backed coaching
In creating Sproutiful, ustwo took the rare approach of bringing together UX designers with behavioural science researchers. Working side by side, we quickly realised just how much science-backed coaching overlapped with the core principles of UX design: be highly empathic, don’t make assumptions, and meet people where they are.
Our first step was to ask what aspects of a real-life coaching experience we wanted to incorporate into the app. As an experiment, I conducted 1:1 coaching sessions with members of the team, with a focus on increasing their veggie intake.
From these real coaching conversations, we quickly realised how crucial it would be for the app to cater to the individual circumstances of users. For example, one team member disliked veggies but wanted to have a healthier pregnancy, while another was vegan and looking to eat a wider variety of vegetables on a budget. It wouldn’t have made sense to address these clients’ different needs in the same way. And yet so many apps on the market do just that.
By creating a custom understanding of what the user cares about most, we were able to bring support and nudges to the user that are actually helpful. For example, if one user wanted to eat more fruit, another wanted more leafy greens, and another user wanted to try new recipes with legumes, each would get coaching nudges that met their unique needs. Such customised nudges are also more effective at re-engaging users who have dropped off – again, increasing adherence.
Looking back on my medical training, I can say that the customised relationship Sproutiful builds with the user is arguably more caring than the relationships many patients experience with their own clinicians. After marrying UX and behavioural science at every step, the Sproutiful coach emerged as a truly unparalleled confidant and gut health guide.
The promise of AI and science-based coaching
With the seemingly relentless pace of AI advancements, many companies are struggling to figure out how they can leverage this technology to better serve users. Sproutiful shows us the power of bringing together the results-oriented approach of behavioural science with the customisation, adaptiveness, and human-like connection of AI. The combination leads to increased adherence, higher engagement, and better results.
This dynamic pairing of AI and science-backed coaching offers benefits to industries beyond health care. By unlocking new levels of personalisation and emotional engagement, it creates the possibility for better digital experiences in any situation where people want to make lasting behavioural changes to improve their lives.
In the end, the lesson of Sproutiful is that if you can design a digital experience that feels like a truly personalised coach—one that adapts to the user’s needs, develops a relationship with them, builds trust, and offers a tailored experience—you’ll have something people want to come back to. The best products aren’t just about solving problems; they’re about creating connections and building trust. That’s what keeps people coming back and sticking with you for the long haul.
* Amagai, Saki, et al. "Challenges in participant engagement and retention using mobile health apps: literature review." Journal of medical Internet research 24.4 (2022): e35120.
Dr. Om Lala is a doctor, behavioural scientist, and coach. He uses science-based coaching techniques to help people change their behaviour and achieve their goals. Dr. Lala taught Harvard’s course on coaching science, led the Institute of Coaching at McLean/Harvard Medical School, and wrote Harvard Business School’s first case on coaching effectiveness.