There’s a comfortable assumption running through the health tech conversation right now: that AI-powered apps, armed with streams of biometric data, are delivering genuinely personalised care. It’s an interesting story. The reality is considerably less impressive.
Most health apps collect a remarkable amount of information, including sleep patterns, heart rate variability, activity levels, nutrition logs, and then produce advice that could apply to almost anyone. “Walk more.” “Improve your sleep hygiene.” “Eat a balanced diet.” The gap between data collected and insight delivered is wider than most users realise, and wider than most investors acknowledge.
The personalisation promise versus reality
According to health and fitness app benchmark data, the average retention rate for health and fitness apps in the UK sits at around 3% by day 30. That means more than 95% of users disengage within a month, not because health goals disappear. But because people quickly sense the advice isn’t meaningfully tailored to them.
The commercial logic driving this is easy to understand. Short subscription cycles, ad-supported models, and low user tolerance for data-sharing friction all push developers toward simple tracking with generic feedback loops. Building clinical depth is slower, more expensive, and harder to monetise quickly. Engagement products get built; clinical tools remain aspirational.
What real data-driven care demands
Meaningful personalisation looks nothing like a canned dashboard tip. It requires longitudinal, cross-provider data and primary care records. It also requires hospital history, pharmacy data, and genomic information, integrated into decision-making rather than siloed in a single app’s ecosystem. This is precisely the kind of infrastructure that takes years to build and governance frameworks to protect.
Digital sectors beyond health have grappled with similar questions about what genuine personalisation requires. Players using online casinos in the UK, for instance, encounter platforms that adapt interfaces based on real-time behavioural signals, a reminder that data-responsive experiences are achievable.
In ecommerce, online retailers continuously refine product recommendations, search results, and promotional offers based on a few factors. This includes browsing habits, purchase history, and engagement patterns to create highly tailored consumer experiences.
This is only possible when the underlying data architecture is built deliberately to serve them. Health tech demands an even higher standard: clinical validation, audit trails, and clinician involvement, not just marketing-grade “AI” labels.
Where behavioural data sets a higher bar
Behaviour is where the gap becomes significantly visible. Research comparing AI-driven chatbot responses to those of human therapists found that chatbots frequently delivered generic replication-style advice rather than context-specific guidance.
This is exactly the problem that health tech companies claim their products solve. When a chatbot gives the same response to markedly different user profiles, the “personalisation” label becomes difficult to defend.
This matters because users increasingly sense when they’re receiving boilerplate. Drop-off rates reflect not just a lack of habit formation but a reasonable judgment: the app isn’t earning continued engagement because it isn’t adding contextual value.
The gap in health tech still needs to close
Promising developments exist. NHS App redesigns and AI-powered triage pilots show what integrated, longitudinal data can achieve when governance is taken seriously.
As research into secure health data sharing makes clear, meaningful personalisation depends on users trusting how their data is used, and that trust is built through transparency and genuine control, not consent tick-boxes buried in onboarding flows.
The health tech sector has built an impressive data collection infrastructure. The harder work, translating that data into advice that accounts for individual risk profiles, comorbidities, and medication interactions, is still largely unfinished. Until the industry closes that gap, most health apps will remain better at tracking lives than improving them.
