BeeSmart is an AI-powered mobile app designed to help beekeepers — especially those just starting out — manage Varroa mite infestations and track hive health over time. It functions as a digital co-pilot: part intelligent assistant, part long-term logbook.
Varroa mite infestation is one of the most serious threats to bee colonies worldwide. Managing it correctly requires consistent monitoring and timely treatment. The challenge was designing an app that genuinely helps someone who has never kept bees, without being condescending to someone who has kept them for twenty years.
The challenge
Beginner beekeepers often can't recognise early warning signs of infestation or know how to interpret the data they collect. Existing tools are either too complex or too generic to be useful at the moment of decision — during an inspection, outdoors, often with gloves on.
The dual-audience problem made this particularly demanding: design too simply and you lose experienced beekeepers who want raw data and quick entry; design too densely and you abandon the beginners who need it most. Most apps resolve this by just picking one audience. We wanted to serve both.
How we approached it
The project began with user research — interviews with beekeepers at different experience levels and field observations during actual hive inspections. Watching people work with their hives in real conditions was essential: it revealed what information they reach for in the moment versus what they reflect on later, and how the physical context of the work (outdoor, gloves on, time-sensitive) shapes every design decision.
Key insights:
- Beginners need guidance at the moment of inspection — after the fact is too late
- Experienced beekeepers want data history and trends, not explanations they already know
- Both groups need fast, friction-free data entry during hive work — outdoors, gloves on, no time to navigate
- The app needs to adapt its depth to the user without asking them to configure it manually
- Outdoor use demands high contrast, large touch targets, and minimal text input
What we built
The final prototype covers the core workflows: users input infestation counts and receive AI-guided treatment recommendations with timing and dosage; every inspection is logged automatically, building a long-term dataset per hive; multi-year charts show infestation patterns over time to support data-driven decisions.
The adaptive UI is the structural solution to the dual-audience problem: novice mode surfaces contextual guidance and plain-language explanations; advanced mode prioritises raw data and quick entry. The switch happens based on declared experience level and usage patterns — not a manual toggle buried in settings.
- AI treatment guidance: Input infestation counts, receive specific recommendations on timing and dosage
- Hive logbook: Every inspection recorded automatically, building a per-hive long-term dataset
- Trend visualisation: Multi-year charts showing infestation patterns over time
- Adaptive UI: Novice mode with guidance vs. advanced mode with raw data — no manual configuration
- Outdoor-optimised design: High contrast, large touch targets, minimal text entry throughout
BeeSmart showed how rigorous user research translates into design decisions that wouldn't have been made without it. The dual-audience structure — designing for beginners and experts simultaneously without making either feel like an afterthought — is the kind of problem most apps sidestep. Solving it required the research to tell us which decisions belonged to which user, and the discipline not to simplify our way out of the complexity.
It also reinforced how AI adds genuine value in niche domains when the prompting and context are done thoughtfully — not as a feature, but as the thing that makes the product useful.