Retail
AI-powered retail management system
Smart shelf management
About The

Project

A computer vision model wrapped into an application for supermarket merchandizers, capable of recognizing packages of particular goods and brands on the shelves, analyzing the amount in stock, the layout, and the popularity per brand. In this way, the AI/ML model helps the merchandizers and supermarket management understand the possible gaps in their strategy and alert via push notifications whenever possible actions can be suggested for increased efficiency.

.NET

MongoDB

Identity Server

Angular

Xamarin

About The

Client

The client is an aspiring mid-tier supermarket and mini-supermarket chain operating in Central and Northern Europe in a highly competitive environment. The competitive advantage embodied in their business model involved not just product diversity and quality, but also the need to constantly upgrade their merchandizing capacities and logistics, leading for them to adopt an AI model to handle the latter aspect of their operations.

Location:
Central & Northern Europe
Partnership period:
2021-2024
Industry:
Retail
OUR

Business Challenge

Our team was tasked with creating a scalable YOLO-based computer vision model wrapped in mobile/web UI for internal use that would streamline and automate product recognition on the shelves, be integrated with the existing digital infrastructure, and allow for real-time reporting and notifications to the responsible users.

OUR

Project Goals

1

Recognize the type of product, package, brand, and amount on the shelf through computer vision

2

Analyze the popularity of each brand and sub-brand based on the available information

3

Monitor the amounts of products in stock

4

Monitor and keep track of sales per product based on adjustable criteria

5

Visualize the data and provide push notifications to merchandizers in certain scenarios

6

Provide real-time reporting and heatmaps

7

Integrate with the existing ERP and other systems

Results

Lionwood.software implemented the computer vision model to process visual information (obtaining that information could be “crowdsourced” among employees, reducing the need for manual tasks to the bare minimum) and AI-based algorithms to provide suggestions and visualize data in real-time.

The main features included:

  • scanning the contents of shelves to obtain the information for the model to process
  • recognition of brands and packages
  • dashboards for analytics
  • push notifications for merchandizers triggered by certain scenarios

The outcome, besides improved shelf management, also included a 25% increase in sales efficiency attested 3 months after system rollout.

Implications

  • Retailers of various business size can use AI to optimize stock levels, reduce out-of-stock instances, and enhance the overall supply chain efficiency
  • Brand visibility and marketing insights can now be affordably collected via computer vision models, aiding in tailoring marketing strategies, optimizing product placements, and improving brand visibility
  • Consumer good companies and brands have a new way of interacting with merchandiser across the distribution networks via AI-enhanced applications
Our projects

Case Studies

View all
HealthCare
2019 - 2022
DS
View
Transportation & Logistics
2022 - 2022
FetchApp
View
Contact us

To find a perfect solution

    Terms of Service
    Privacy Policy
    contact-us-image