RETAIL

There is a paradox at the heart of the digital transformation of retail. On the one hand, AI offers retail businesses some undeniably attractive capabilities. Computer vision product recognition enables self-service checkout, automated restocking, and loss prevention solutions. Behavior recognition allows companies to create personalized shopping experiences for their customers. And behind the scenes, automated data analysis means streamlined operations and better supply chain management.

But on the other hand, many businesses are still wary of AI solutions—even though they recognize the potential benefits.

“There are several reasons why retailers are hesitant to adopt AI solutions, but the biggest factors by far are the lack of in-house technical skills needed to implement them—as well as plain old fear of the unknown,” says Liangyan Li, Head of Global Sales at Hanshow, a solution provider of digital store solutions for the retail sector.

There is justification for such concerns, because implementing AI in a retail setting entails significant technological hurdles. To begin with, it means building a high-performance system that can process vast amounts of data in real time. In addition, there is an innate complexity to retail automation, which usually involves multiple technologies and computing workloads. And finally, there is an element of IT overhead as well: the ongoing need to monitor and maintain a solution after deployment to ensure stability.

The good news for retailers—and for retail systems integrators—is that a new era of ready-to-deploy edge AI solutions has already begun. Built atop next-generation processers, and using software tools designed for edge computing, these solutions offer simple, effective implementations to would-be adopters.

What is the key to building solutions that meet the needs of #retail businesses? The combination of industry-specific #AI know-how and enterprise-tier #technologies designed for ease of deployment and performance at the #edge. @hanshowofficial via @insightdottech

Edge AI Solutions Engineered for Retail

What is the key to building solutions that meet the needs of retail businesses? The combination of industry-specific AI know-how and enterprise-tier technologies designed for ease of deployment and performance at the edge.

Hanshow’s hardware and software technology stack, combined with its experience in developing AI applications for retail, enable a flexible, user-friendly solution—and one that addresses the traditional concerns of business decision-makers in the sector. Here, Li credits Intel with helping to bring Hanshow’s solution to market.

“Intel is unmatched as a platform for stable, reliable edge computing—particularly when attempting to develop a comprehensive, seamless solution for the end user,” says Li.

Hanshow’s solution incorporates a number of different Intel technologies:

• Intel^® Core^™ Processors handle heavy edge workloads and image processing tasks
• Intel^® Media SDK gives developers access to media workflows and video processing technologies—shortening time to market
• The Intel^® OpenVINO^™ Toolkit speeds AI application development and helps optimize visual processing algorithms
• Microsoft Azure Cognitive Services allows developers to build sophisticated AI algorithms even if they don’t have machine learning experience

On a practical level, Hanshow’s Intel technology-based solutions have the added benefit of being relatively easy to implement in a working environment—and can thus bring about dramatic improvements to operational efficiency in a very short time.

Smarter Shelves from Europe to Japan

Hanshow’s smart shelf management deployments in Europe and Japan are case in point.

Despite the geographical distance, both of Hanshow’s retail customers faced similar challenges: a need to gain greater insight into what was going on in their stores to improve efficiency and boost sales.

The European business, a large supermarket chain with a global footprint, was facing frequent shortages of fresh food in its stores. The main cause of this problem was the inability of employees to identify out-of-stock (OOS) products and take steps to replenish them in a timely fashion.

The Japanese company, a large chain of department stores, was having difficulty identifying the habits and preferences of its shoppers, hampering the business’s marketing efforts.

Hanshow implemented a comprehensive AI solution at both companies. In the supermarkets, it used computer vision cameras to take images of fresh food stacks to provide near real-time data on stock. In the department stores, the company implemented a digital shelf solution that encompassed marketing, OOS management, human-product interaction, customer demand analysis, and smart advertising.

The results were dramatic. The supermarkets saw their average OOS duration drop from 2.5 hours to 1.5 hours—a 40% improvement—while also eliminating the need for employees to perform daily manual inspections. The department store chain, for its part, saw an immediate effect on sales: an increase of nearly 20% in sales of active products when single-product recommendations were implemented in digital shelf areas.

The Transformation of Global Retail

The promise of AI in the retail sector is not new. But the emergence of comprehensive, easy-to-deploy solutions will turn that promise into a reality.

It’s hard to overstate the effect this will have—especially as adoption increases, and systems integrators and technology companies begin to develop the retail AI ecosystem in earnest. Expect to see more complex computing workloads, multi-architecture applications, and new benchmarks for operational efficiency and consumer experience.

This is why Li talks in terms of the “transformation of the global retail market.”

“AI helps retailers provide consumers with more personalized services, accelerates business operations and commodity circulation, and delivers more valuable data insights,” he says. “It will allow retailers to reshape the relationship between people, products, and markets.”

Edited by Georganne Benesch, Associate Editorial Director for insight.tech.