Looping
Waste sorting with A.I
Ever stood over the trash, paralyzed by the question: “Is this plastic, or is it just pretending?” Welcome to Belgium, where recycling is basically an Olympic sport and everyone’s losing. Guides? Confusing. Bins? Different by region. Outcome? Guilt and chaos.
My Role
Chief Design Officer
UX Designer
Researcher
Team
Worked cross-functionally with my 2 co-founders + AI Engineers
Objectives and goals
• Make waste sorting instinctive (literally: “Open app. Snap. Sorted.”)
• Swap recycling anxiety for confident everyday habits
• Localize because Brussels ≠ Wallonia ≠ Flanders
• Conduct usability tests with dozens of citizens
• Run joint campaigns, drive virality with influencers
• Create videos and visuals with generative A.I tools
Video presentation of Looping, the Shazam of sorting
Context & Challenge
Mission
Empower every Belgian to sort waste correctly, daily, and stress-free, with no manual required.
Structural Challenges
• Perfecting AI for hyper-local rules (there’s always an exception)
• Onboarding boomers and teens, equally
• Competing for attention in a crowded app space
• Create an easy to use experience for every citizens
Research & Insights
Discovery
• We noticed that 70% of people we interviewed found that sorting was not that complicated. When they tried ou sorting games, they were wrong on many items.
• During our tests, people were convinced that Looping was all A.I, at first it wasn't at all. It was us answering people questions up to midnight.
• We were worried about the delay between the request to an expert and the response. It appeared that it wasn't an issue at all
Screen recording of the Ecoquizz game where people have to select the correct way to sort an object
Problem Statement
How might we make "Belgium’s shazam of sorting" easy for everyone ?
Looping getting famous in the news. Fun fact, our notifications went crazy because of thousands of requests in minutes
Strategy & Approach
Pragmatical approach: People got answers from our team first, not knowing it wasn't AI. It allowed us to gather thousands of picture to train an AI model.
Camera first: App opens to camera, point, snap, answer. No bullshit
Localization engine: Automatically adapts to city/region rules based on zipcode
Instant corrections: User can choose between 3 automatic suggestions for identified objects
Gamification: Launch of a mini game inside the app to test users knowledge
In-app tips: Tips appear on products that have more durable alternatives
Kiosk: We decided to create a Kiosk version of Looping so compagnies could help their employees to be properly educated on sorting
Our kiosk helping Decathlon teams
Design & Prototyping
I designed around 15 versions of the sorting rules screen, we released several test versions to track the reaction of citizens
We tested the app with friends, family, colleagues, influencers & large office employees to evaluate reactions and gather feedback
At first, we were not sure about the value of our app so we decided to add feedback buttons to measure satisfaction and frustration.
How to recognize waste with AI by snapping a picture with Looping
Impact
Thousands of Belgians onboarded within months without any marketing
Error rates in sorting down (users: “I can finally trust the answer!”) measured with our game results
Identify which packaging causes sorting issue and what can we do about it
89% of our users found that Looping was helpful
4.8/5 global rating score on appstores (800 ratings+)
Partnership with big brands like Carrefour, Danone and Decathlon
Be Circular selected project 2023
Reflection & Learnings
Biggest surprise: Citizen were confident about their skills at sorting but they failed at our game, demonstrating a real difference between what they know and what they think they know
Hard lesson: Think fast but ship faster, AI is evolving at incredible speed, so keep yoursef updated every week
What matters: Be pragmatical, clarity, simplicity and ease beat innovation-for-innovation’s-sake, every time.