RSS
Логотип
Баннер в шапке 1
Баннер в шапке 2
Project

VkusVill began using an AI system in stores that recognizes spoiled fruits and vegetables

Customers: VkusVill

Moscow; Trade

Contractors: Art Intelligence
Product: Artificial intelligence (AI, Artificial intelligence, AI)

Project date: 2024/07

On July 31, 2024, the herd is aware that an innovative system based on artificial intelligence has been introduced in the VkusVill chain of stores to determine the quality of fruits and vegetables. A system based on neural networks allows you to quickly recognize product defects, preventing poor-quality goods from entering shelves.

The technology is developed in collaboration with technology startup Art Intelligence and is already used on more than 60 darxtors of the network. The system is integrated into the work of darkstore employees, who select fruits and vegetables daily for write-off or markdown in the event of defects. The AI system helps to significantly speed up this process by providing feedback on each request in a matter of seconds.

source = TasteWill
VkusVill began using an AI system in stores that recognizes spoiled fruits and vegetables

The basis for the functioning of the new system was computer vision algorithms and a neural network trained in photographs from the history of chat rooms of darkstore employees. Interaction with AI takes place through a chatbot in the Telegram messenger, where employees send photos of products. The chatbot responds by saying whether to send the product to the "green price tag" or write it off, and also provides a brief help on the identified defects.

As the innovation manager of VkusVill Nikolay Belyaev noted, the technology for recognizing defects using neural networks is not new, but its use to assess the quality of fruits and vegetables is an innovative approach. The project team deliberately abandoned strict requirements for the quality and positioning of objects, which made the technology available for everyday use by employees.

For July 2024, the system recognizes defects on products such as citrus, apples, bananas, courgettes, tomatoes and cabbage. The project team plans to expand the list of products, including exotic fruits, and improve algorithms to indicate specific defects. In the future, it is planned to achieve complete process automation with the ability to integrate the system into any internal business processes and resources of the company.