Lenta Integrates AI into Fresh-Produce Quality Assessment
Computer vision and neural networks are enabling Lenta Group to expand remote inspections of its most vulnerable fresh‑produce categories, helping the retailer predict defects earlier, reduce waste, and improve food quality for millions of consumers.
Remote Quality Control
Retail chains are required to sell only safe, high‑quality food products, making rigorous incoming quality control a critical operational task. The system must prevent spoiled or damaged items from entering distribution or deteriorating during storage and transit. Because this labor‑intensive process demands continuous human attention, retailers are increasingly adopting automated inspection workflows.
Lenta began this transformation in 2021 by launching remote acceptance of fresh produce at its distribution centers. These items are among the most fragile: they can be damaged at any stage of the logistics chain, whether at the grower level or during transportation. Incorrect packaging that bruises produce or slight deviations in cold‑chain conditions can rapidly degrade product integrity.
At high delivery volumes, identifying every defect is extremely challenging. In 2021, Lenta established remote decision‑making centers staffed with agronomists and qualified quality specialists. Facilities were equipped with dedicated inspection tables, enhanced lighting, and high‑resolution cameras. Inspectors assess incoming produce using headsets, large monitors, and joystick‑controlled cameras. Each specialist can review more than six tons of fresh produce per month.
The process continues to evolve. Artificial intelligence is now being integrated into the workflow.
Computer Vision Meets Neural Networks
Because fresh‑produce inspection is highly repetitive, it is well‑suited for automation. Lenta has begun testing a computer‑vision system powered by neural networks. Early results are strong: in over 90% of cases, the system’s assessments align with those of experienced inspectors — a promising signal for scalable deployment.
Digitization reduces cognitive load on staff and accelerates inspections and batch processing severalfold. It also standardizes evaluation criteria, minimizing human variability caused by fatigue, shift differences, or subjective interpretation. Large‑scale AI deployment in incoming inspections is planned for mid‑2026.
Lenta’s next digital milestone is hyperspectral imaging, capable of identifying produce ripeness. The camera captures the reflected‑light spectrum for each pixel. Using this data, Lenta’s innovation team applies AI models to analyze peel and tissue characteristics — pigments, moisture, lipids, proteins, and other chemical parameters. The resulting “spectral fingerprint” correlates with flavor, juiciness, and shelf life. The technology is currently being tested on citrus fruits and will later expand to other produce categories.
Prospects
“One of our priorities for the coming years is deep automation and the integration of artificial intelligence,” said Yuliya Batenina, Lenta Group’s quality director. “We aim to deploy computer vision for assessing fresh‑produce quality and supporting production‑line staff. We are also launching a major project to monitor temperatures across the entire supply chain — from warehouses and transportation to in‑store refrigeration equipment.”
In the long term, these technologies may become integral to comprehensive smart supply‑chain systems — from farm to shelf — with end‑to‑end visibility into storage and transport conditions. Such visibility increases transparency, reduces food waste, and improves product quality for consumers.
The system can also integrate with Russian smart‑farming platforms, enabling retailers to monitor the full lifecycle of agricultural products: production, harvesting, transport, storage, and delivery. This ensures that quality standards are met at every step.
Once fully deployed and validated in Russia, the technology may support exports of Russian food products by serving as a digital quality certificate that verifies compliance with standards. Russian developers involved in such projects could offer similar solutions to international partners, particularly in markets where continuous quality verification is a regulatory requirement.
