WHAT CAN I EAT

As someone who does not always have a direct answer to the question "What do you want to eat?" I decided to build an app that helps me answer this question quickly and with useful options.

Refining the problem

The question of what can I eat is one that comes in a variety of use-cases, and to build a solution that properly met the needs of people that had the question, I had to capture all scenarios possible. I distilled them to three use-cases.

• Knowing you wanted a specific flavour or texture but not knowing a specific food

• Wanting to make something (whether new or old) with ingredients at home

• Avoiding the question by having food already aligned to your palate available

Building with AI

This was the first product I built end-to-end using AI technology. I began by creating a product specification in plain language which governed the project.

• I proceeded to define the visual identity and design system, choosing a simple and fun theme (because food should be fun) I went with a playful font and green accents.

• I then built the architecture of the site, and refined journeys using UX principles (like localisation, personalisation, error avoidance, error recovery, plain and simple language, and clear navigation and sign-posting) to ensure it was an accessible and useful solution.

• I created a lot of iterations to ensure the resulting solution met user needs.

Building with AI allowed me to design and develop this tool significantly faster than was ordinarily possible.

Designing for scale and complexity

Scalability was a core design principle from the outset. I ensured this solution worked for 10 users as well as it worked for 10,000 by implementing a low entry barrier with seamless onboarding, and personalisation options that allowed users save their recommendations, meal plans and recipes. I also implemented safeguards that ensured the tool was not retaining unnecessary information like fridge images after scanning causing system bloat. I also took appropriate security measures to avoid exploitation by bad actors.

Evidence based design

As users began to try the app and give feedback, I gained valuable insights which I incorporated. One example is the implementation of checkboxes on grocery lists generated for meal plans and the ability to name meal plans.

Key features

• Craving finder: Users can swipe through craving suggestions based on flavour and texture with results linking to cooking recipes and takeout options.

• Fridge scanner: Users can scan their fridge, cupboard or pantry to get written or video recipe recommendations to follow along.

• Meal planner: Users can generate meal plans for various 1, 3, 7, or 30 days and get a corresponding grocery list.

• Multi region support: Meal recommendations take user location into account o ensure relevance and availability and the platform supports 12 languages.

  
  

Outcomes

A user friendly, easy to use web app that provides meal recommendations.