MmaDiThini

MmaDiThini– Banking on Waste, Banking on People.A smart recycling system that transforms everyday waste into income and inclusion for underserved communities.

Comment

Inspiration

MmaDiThini was born from a desire to transform Botswana’s growing waste crisis into an opportunity for financial inclusion, digital transformation, and eco-conscious innovation. Inspired by the informal waste sector, youth tech adoption, and FNB’s commitment to sustainability, our goal was to build a smart reverse vending system that rewards positive environmental behavior. We envisioned a future where a recycled bottle could be a bank deposit, giving people—especially those in the informal economy—a chance to participate in both the green and digital economies.

What It Does

MmaDiThini is an AI-powered, gamified recycling machine linked directly to the FNBB mobile app. Users deposit a recyclable (plastic bottle, metal can, or glass), and the machine:

Detects and classifies it using YOLOv8 object detection. Sorts and compacts the material via conveyor and servo-driven systems. Updates the user’s FNBB-linked MmaDiThini account with reward points. Rewards can be redeemed for airtime, shopping vouchers, electricity, or even FNBB wallet top-ups. Includes a leaderboard, and badges.

How We Built It

We combined hardware and software components into an integrated system:

Hardware: Built on ESP32 and Raspberry Pi, with IR sensors, a Pi camera, motors, and a steel housing for durability.

Software: YOLOv8 trained with custom datasets for object recognition. Python-based classification enhanced with heuristics (e.g., brightness, shape). n8n AI agent automates rewards, data logging, and system alerts.

FNB App Integration: Users track activity, manage rewards, and interact with eco-campaigns via the FNBB app. Custom FNBB-MmaDiThini bank cards for financial onboarding and spending.

Challenges We Ran Into

Hardware Integration: Syncing sensors, object detection, and motor control required extensive testing.

Data Accuracy: Training the AI model to distinguish materials under varied lighting and shape conditions was complex.

Scalability: Designing a solution that works both in malls and rural areas needed modular, low-power hardware and solar compatibility.

User Behavior: Changing mindsets around recycling meant we had to make the system engaging, not just functional.

Accomplishments That We’re Proud Of

Successfully deployed a working prototype capable of real-time detection and classification.

Demonstrated a seamless reward integration with the FNBB mobile ecosystem.

Aligned our platform with FNBB’s strategic pillars: ESG, digital banking, and financial inclusion.

Created an impact model that goes beyond waste—it banks behavior.

What We Learned

Gamification is a powerful motivator—leaderboards and instant rewards drove more interest than static recycling bins.

AI is effective even on constrained devices when optimized well.

Financial inclusion can begin with something as small as a bottle—behavior-based onboarding is a real pathway.

Community engagement and partnerships (with municipalities, malls, and FNBB) are key to adoption and scale.

What’s Next for MmaDiThini

Pilot Launch: 10 machines deployed in Gaborone, Francistown, and Maun.

Deep FNBB Integration: Real-time wallet crediting, carbon tracking dashboards, and co-branded cards.

Scale to 100+ units over 3 years with CSR partnerships.

Introduce eco-credit scores, giving users a financial footprint based on positive actions.

Expand into other SADC countries as the first smart recycling bank-led platform on the continent.

Built With

Share this project:

Updates