MuseMeal: Supporting Healthy Cooking through Embodied Conversational Interaction and Wearable Projection
Overview
Developed as a project submission for the ACM DIS Student Design Competition 2026, MuseMeal is an AI-powered healthy eating companion app designed to help university students build sustainable cooking habits.
Services
UX & Research
Year
2025-2026
Link
01- Context
The problem isn't knowledge, it's support.
Eating healthily is hard for students living independently for the first time. Time pressure, tight budgets, and emotional exhaustion push them toward convenience food — and the gap isn't that they don't know what a good meal looks like, it's that nothing supports them in the moment they're deciding, shopping, or cooking. We anchored the project to SDG 3 (Good Health & Wellbeing) and SDG 12 (Responsible Consumption & Production).
02- Problem
Existing healthy diet tools tend to only tend to track calories and foster good meal planning. They assume motivation, planning capacity, and energy that a student at the end of the day simply doesn't have. The decision burden of "what do I cook, with what I have, that's healthy" goes unaddressed; and the cooking itself is a lonely, unguided task with no reward for getting it right.
03- Process
User Research & User-Centred Design Iterations
We grounded the design in a survey (n=46), 10 semi-structured interviews and testing with 20+ users. We moved through user-requirement definition, brainstorming, and structured idea evaluation against impact, accessibility, usability, and feasibility before building a low-fidelity prototype. Two key decisions shaped the final system:
Decision 1 — meet the user's mood, not just their pantry. Rather than a static recipe search, we built a generative AI assistant, Radish, that takes current mood, taste preferences, and available ingredients as input and returns personalised options using an AI algorithm.
Decision 2 — pivoting from projection to a smartwatch companion. Our first concept used a projector-based system; user testing showed it was clunky and impersonal, so we shifted to a hologram-style avatar fixed to the smartwatch face, giving simplified step-by-step audio-visual guidance and encouragement hands-free while cooking. This made the guidance more personal, portable, and emotionally engaging.
04- Solution
Two systems, one habit loop
The final design has two core components. An AI-driven avatar system generates personalised recipes through Radish and then coaches the user through cooking, step by step, via the watch-face companion. A reward system closes the loop: users photograph the finished dish to earn points redeemable for discounts on ingredients or healthy meals at partner stores, tying home cooking to tangible real-world value. A second iteration refined the interaction flows and visual language based on feedback, making the UI warmer and clearer.
05- Outcome
Published, and validated with real users
MuseMeal was published and presented at ACM DIS 2026 as part of the Student Design Competition, and was awarded a place in the top 4 projects. The design was grounded in and tested with real users across three iteration cycles.
