ideas are brewing.

not quite...

← Back to work Haleon · 2022 – 2023

Consumer Health
Intelligence Platform

Role Principal Product Designer
Scope 0→1 product design
Team 3 designers · 2 PMs · 22 engineers
Platform iOS · Android · Web
Product overview — personalised health dashboard

Personalised health guidance at global scale

Haleon is one of the world's largest consumer health companies, home to brands like Sensodyne, Panadol, and Voltaren. Despite having hundreds of millions of customers, their digital touchpoints were generic — the same content served to everyone, regardless of health profile, geography, or product history.

I was brought in to design a 0→1 AI-powered platform that could personalise health recommendations at scale — surfacing the right product, advice, and educational content for each individual, while staying within strict regulatory boundaries across 20+ markets.

38%
Increase in repeat engagement with health content after personalisation
2.4×
Higher product recommendation conversion vs. generic content
22
Markets launched within 12 months of design handoff

Personalisation without overreach

Consumer health is a uniquely sensitive design space. People share deeply personal information — symptoms, medications, chronic conditions — and expect it to be handled with discretion. Any system that felt surveillance-like would erode trust instantly, and in regulated health markets, the line between advice and medical recommendation is both important and legally significant.

Our research revealed a consistent tension: users wanted the platform to "know them" but were uncomfortable with anything that felt like it was making inferences beyond what they'd explicitly shared. This became the central design problem — how to be genuinely helpful without being presumptuous.

Research — user interviews & trust framework mapping

A progressive disclosure model for AI-assisted health

We designed around a concept we called "earned personalisation" — the platform starts generic and becomes more specific only as users actively share context. Every step toward personalisation is explicit, reversible, and explained in plain language. There are no dark patterns, no inferred health states without consent.

The AI recommendation engine was designed with a clear visual language distinguishing three types of output: educational content (always available), product suggestions (shown when relevant), and symptom-specific guidance (shown only after user consent and with a clear disclaimer). This hierarchy shaped the entire information architecture.

We ran co-design workshops across five markets — UK, US, Germany, Brazil, and Australia — to pressure-test assumptions about trust, health literacy, and cultural attitudes toward digital health tools. The resulting design system was flexible enough to localise without fragmenting the core experience.

Health guidance people actually return to

The platform launched in five pilot markets and expanded to 22 within the first year. Repeat engagement with personalised health content was 38% higher than with generic content in A/B testing. Critically, opt-in rates for deeper health profiling exceeded projections by 60% — a signal that users trusted the transparency model we had built.

The design system and "earned personalisation" framework were adopted as Haleon's global standard for all digital health experiences, informing product strategy across the brand portfolio.

Next project

NHS — Clinical Decision Support →