Human or AI Coaching Keeps People Using Health Apps, Review Finds

Health apps are handy, but let's be honest: it's easy to download one, poke around for a day or two, and quietly ghost it by Thursday. So what actually keeps us showing up? A systematic review points to one promising ingredient, coaching, and asks a timely question: does it matter whether that coach is a human, an artificial intelligence, or some blend of the two?

What the researchers wanted to know

Digital health interventions, or DHIs, have been championed as a scalable way to promote health and prevent illness, aligned with global goals like the United Nations Sustainable Development Goal for good health and well-being. But they face a persistent criticism: they are often criticized for "shallow and transactional engagement," and keeping people using them over time is hard.

The reviewers zeroed in on a potential fix, pairing these digital tools with health coaching, which brings a more meaningful, human touch. Their goal was to synthesize the existing peer-reviewed research on coach-facilitated DHIs: to understand how coaching is actually being used inside digital health tools, and what impact it has on both engagement and lifestyle outcomes.

How they studied it

This was a systematic review, a rigorous method for gathering and appraising all the relevant research on a question rather than cherry-picking studies. The reviewers included studies that examined DHIs with a coaching component addressing lifestyle outcomes. They searched four major academic databases, APA PsycINFO, Medline, Web of Science, and Scopus, from each database's inception up to February 2025.

Three authors handled the study selection, appraised the quality of the studies using a recognized tool called the Mixed Methods Appraisal Tool, and extracted the data, capturing details about study characteristics, coaching features, participant engagement, and lifestyle outcomes. The findings were then brought together through a narrative synthesis, weaving the results into a coherent overall picture.

What they found

Thirty-five studies made the cut, and from them the reviewers identified three distinct coaching modalities inside digital health tools: digital human coaching, artificial intelligence (AI) coaching, and hybrid coaching that combines human and AI. Encouragingly, all three modalities "demonstrated feasibility and acceptability," meaning they were practical to deliver and well-received by users.

Both human coaching and AI coaching showed a "positive impact on both engagement and lifestyle outcomes," a notable result given how central the engagement problem is to digital health. The hybrid approach, which aims to blend AI's scalability with the depth a human can provide, was seen as promising but in need of further refinement to fully realize that combination.

What this means for you

If you've ever struggled to stick with a health or wellness app, this review offers a practical clue: some form of coaching may be part of what helps you stay engaged and actually follow through on healthy changes. The reassuring news is that both human and AI coaching were linked to better engagement and lifestyle outcomes, so you don't necessarily need constant access to a human professional to benefit, a well-designed AI coach may help too.

When choosing a health app, it's worth looking for one that offers some kind of coaching or guidance rather than leaving you entirely on your own, since a purely transactional tool is exactly the kind this review found easy to disengage from. And if an app blends human and AI support, keep in mind that hybrid models are still being refined, so expect some variation in how smoothly they work.

It's also worth being honest with yourself about what kind of support actually keeps you going. Some people thrive on the accountability and warmth of a human coach; others prefer the always-available, judgment-free nature of an AI that's happy to check in at midnight. This review suggests both can genuinely help, which means you can choose based on what fits your temperament and your budget rather than assuming one is inherently superior.

The deeper takeaway is that engagement isn't purely a matter of willpower. If you've blamed yourself for abandoning app after app, this research offers a kinder explanation: many tools are simply built to be transactional, and pairing them with real guidance is part of what turns a good intention into a lasting habit.

The honest caveats

A systematic review is only as strong as the studies it gathers, and this one synthesized 35 studies through a narrative approach rather than crunching all their results into a single number, so it summarizes the state of the evidence rather than delivering one precise effect size.

Terms like feasible, acceptable, and a positive impact on engagement are encouraging, but they describe general patterns across varied studies, not a guarantee that any particular app will work for any particular person. The review itself flagged that hybrid human-AI coaching needs further refinement, a candid acknowledgment that this fast-moving area is still maturing.

And because digital health tools and AI capabilities are evolving quickly, findings drawn from research up to early 2025 will keep needing updates. Read this as a well-grounded signal that coaching helps digital health tools stick, while remembering that the details of how best to deliver it are still being worked out.