Score Partners With Two-a-Day To Deploy Vision AI Across Fruit Production

Two-a-Day Group, one of Africa’s largest fruit-growing, packing, and marketing organisations, has entered a research and technology partnership with @webuildscore to expand the use of Score's Vision AI across its production and packing operations.

The partnership focuses on advancing real-time computer vision systems used throughout Two-a-Day’s high-volume agricultural environment. Two-a-Day has long relied on computer vision for fruit sorting, grading, and starch-level analysis, supporting quality control across millions of cartons of apples and pears each year.

As production scales and export standards become more demanding, the company is now focused on extending those systems to improve precision, resilience, and operational efficiency.

By integrating Score’s Vision AI models, Two-a-Day aims to enhance existing workflows while introducing new layers of operational and quality intelligence.

The collaboration is expected to support more advanced defect detection, finer-grain quality analysis across multiple fruit varieties, early identification of yield or quality risks, real-time detection of line stoppages and bottlenecks, and automated reporting with end-to-end traceability from intake to finished cartons.

Score’s models are designed to operate under the variable conditions typical of agricultural production, including inconsistent lighting, diverse fruit sizes and colors, changing conveyor speeds, and seasonal shifts. The objective is to evolve existing tools into a broader system that delivers faster feedback for operators, reduces reliance on manual supervision, protects against quality drift, and provides deeper insight into production trends.

Peiter Uys, computer vision and software engineer at Two-a-Day, said the collaboration reflects the company’s need to continually expand its technical capabilities to remain competitive in global agriculture. He noted that Score’s models strengthen existing processes while enabling new applications across the production environment.

"It’s an important step toward maintaining our quality leadership at scale.”

Maxime Sebti, CEO of Score, said the partnership builds on a foundation Two-a-Day has developed over years of applying computer vision, with a focus on improving accuracy and expanding operational intelligence where consistency and quality are critical.

"This is where modern AI can make a real difference: in environments where consistency and quality truly matter.”

Valuable additional context was shared by @bittingthembits, who pointed to the deployment as evidence of real-world adoption of Bittensor. He emphasized that Two-a-Day is an industrial agriculture company operating at scale, not a crypto-native organization, and that it is actively deploying Vision AI powered by Bittensor through Score’s subnet.

"Bittensor isn’t a demo network. It’s running inside 77-year-old companies that move millions of physical goods. That's adoption. $TAO is infrastructure that gets chosen when failure isn't an option."

Most importantly, Andy highlighted that Score’s models are already delivering real-time defect detection, quality grading, early warning signals for yield and quality risks, line stoppage and bottleneck detection, and full intake-to-export traceability in production environments known for being challenging for AI systems.

• 76.9% accuracy in 4 weeks
• 2.1% from human expert level

Disclaimer: This article is for informational purposes only and does not constitute financial, investment, or trading advice. The information provided should not be interpreted as an endorsement of any digital asset, security, or investment strategy. Readers should conduct their own research and consult with a licensed financial professional before making any investment decisions. The publisher and its contributors are not responsible for any losses that may arise from reliance on the information presented.