Artificial intelligence and Edge computing bring tangible benefits to customers and retailers

In the past years, online shopping has quickly developed. Accessible anywhere, with low delivery cost, online shopping is appealing and convenient for a significant portion of the population. Online shopping has quickly made profit of customer profiling and targeting techniques, hence further improving communication and orientation of potential customers. This trend has accelerated with COVID-19 and associated lockdown measures. In November 2020, it was threatening the survival of retail stores who must constantly adapt to improve their sales.

The first option for retail stores is then to propose online versions of their stores and associated delivery services. The second one is to optimize their local stores to increase benefits and lower costs.

The development of artificial intelligence and data analytics is offering today new ways for retail stores to develop their business. With the combination of IoT sensor information and Video Analytics, deep learning solutions are now able to analyze and interpret different situations.

Edge computing consists in local computing resources, optimized for Artificial Intelligence and directly integrated into the store. Edge computing is therefore ideal for retails stores to process and analyze in real time the vast amount of the video data and IoT sensors captured on site. Remote cloud connection is then only used to aggregate added value information and potential notifications.

The combination of AI deep learning for video analytics and edge computing is therefore supporting the development of the performance of the retail store.

Here are some sample use cases:


Optimize customer experience

To develop customer experience, video analytic solutions can now process information to identify customers itineraries, identify queue delay and abnormal situation or behavior. More specifically, customer itineraries help retailers to optimize shopping experience and increase sales opportunities. This information is crucial for brands, which invest a lot in promotions to boost sales.


Enhance security in the shop

Retailers can increase revenues with broader opening hours up to 24/24 7/7 without increasing staff. Computer vision on edge computing can guarantee security and prevent theft at automatic pay stations thanks to real time video analytics. In case of emergency or abnormal situations, an action is triggered in real time. This allows customers to enjoy shopping at any time of day and night.


Optimize stock management

The use of Computer vision at the edge enables optimization of stock management and inventory. Shortages, overstocking and inventory shrinkage contribute to a loss of 1 000 billion dollars to worldwide retailers every year. To do so, cameras just scan the stock, using drones for example. Deep learning models are then able to count the elements in stock, identify wrong placement, and automatically initiate stock provisioning if required.

The potential of valuable data in physical stores is huge. However, IoT analysis through private and public cloud can be a burden for retailers. Connectivity is often unstable and unequal in every area of the shop floor and in the warehouse.

Edge computing addresses this fundamental problem to harness the full benefits of retail IoT: ensure real-time analysis regardless of the connectivity. IoT data from sensors and cameras is sent directly to the edge server, where deep learning AI algorithms analyze data without latency. The pre-processed data can be sent later to the cloud or the datacenter for re-modeling.

In terms of cost, edge computing can also make the difference as artificial intelligence requires significant compute capacity. Having tailored computing resources on site, at the edge of the cloud is today a lot cheaper to equivalent resources in the cloud. Also, huge network bandwidth and consumption would be required to transmit all the video streams to the cloud. Again, local Edge processing is cutting that cost.

Will retail stores continue to suffer from online shopping? Most certainly, but with the advent of AI and Edge computing, they now have new ways to develop business and ensure customer satisfaction and loyalty.


Learn more on BullSequana Edge

Learn more on Edge computing for retail

Share this blog article

About Damien Faure
Global Head of Edge & AI, Atos

Damien Faure is Global Head of Edge Computing and AI for Big Data & Security at Atos. Damien joined the company in 2015 as head of private cloud engineering. He has directed the development of the cloud portfolio of Atos until 2019. Previously, Damien held different positions at Bull, first leading software development and later on driving data centers and IT services as site CIO. Damien holds a M.S. in computer science and applied mathematics from Grenoble-Alpes University (Grenoble INP – ENSIMAG) in 2000. He also studied various management curricula with EMLyon, Cambridge and Harvard business schools. He was nominated to the “gold for experts” program in 2015 and is also senior expert in the Atos expert community.
Follow or contact Damien