Project SpinRetail by Spindox Labs | Year: 2017 – 2018Ăą

The innovative concept for large-retail organizations is called SpinRetail: the supermarket driven by artificial intelligence.

SpinRetail, the supermarket design of the future is made in Italy, was developed in our research hub in Trento. It’s an articulated concept that integrates different sensor systems and a powerful deep learning engine. There are several touch points between the Spindox solution and Amazon Go, the new cashierless store designed by global giant Jeff Bezos’s engineers. SpinRetail responds to a multitude of objectives: theft prevention, monitoring and optimization of the visitor flow in the points of sale and providing information and advice about the products the consumers are about to buy.

Solution

SpinRetail integrates three main modules (data collection, analysis, reporting) and two front-end applications: one for the store manager and another one for the consumer.

Data collection (Sensors)

SpinRetail applies the logic of sensor fusion: a combination of data deriving from disparate sources. In particular, the solution uses weight sensors positioned on the shelves and beacons for locating people.

Data Analysis (Deep Learning)

All data, collected thanks to the sensors, is conveyed within a cloud application. Deep learning models allow to reconstruct the customer’s path within the supermarket, the products being taken from the shelves and the number of products available in the store.  Maybe add more text here about what is Deep Learning and AI…for the person who does not know the scope and definitions…FYI.

Reporting

SpinRetail generates reports in real time. The store managers can therefore evaluate the situation at any moment: the availability of products on the shelves, the movements of the customers and eventual theft attempts. All information is subsequently saved in a storage system.

Benefits

Theft reduction: thanks to SpinRetail’s ability to monitor the products the customer takes from the shelves, it’s possible to prevent theft inside the points of sale.

More efficient replenishment: the monitoring of the picking of products from the shelves and the generation of automatic out-of-stock alerts provide a real-time overview about the needs of replenishment and consequently a reduction of losses and inventory.

Optimization: knowing how the customers move within a store allows the positioning of goods in a targeted manner and in line with product marketing and communication logics.

Customer support and loyalty: through applications, downloadable for smartphones or tablets, the customer receives promotions and purchase advice before and during the shopping. This way a continuous interaction between consumer and store is being established.

Customizable interface: the web application available for the store manager is fully configurable. The user can customize any type of reporting generated.