Smart Retail: Know Your Customer, Drive Agility in Operation

Artificial Intelligence is reshaping retail store business models in various essential business sectors including data-driven operation and customer analysis, computer vision-powered in-store autonomous experience, and loss prevention. After over 2 years of the pandemic, retail business owners meet new challenges as customers are expecting new and improved experiences when they get back to stores. From our partner Tryolabs’s takeaways from Shoptalk 2022 in Las Vegas, Marie Driscoll from Coresight mentioned:

Consumers expect everything, everywhere. They expect digital access in the store; functionality quick. Sometimes you go in wanting a quick, efficient pick-up; sometimes, you want to be romanced. So stores have to be efficient and sexy.

Tryolabs: The physical store is back
The state of retail in 2022: takeaways from Shoptalk

Driven by AI technology, business owners start to use prescriptive and predictive analytics, rather than based on tradition, experience, or intuition. Real-time analysis and automation will help retail business organizations build a trained and trusted workforce in both in-store optimization and back-store operation. Mckinsey believes the Store of the Future is likely to achieve EBIT(Earnings before interest and taxes) margins twice those without automation, with the improved customer experience, better employee engagement, and an easier-to-run store.

‘Future’ co-workers have arrived in retail stores. 

In-store optimization: 

Customers analysis:

Sentiment analysis: Zenus’ smart camera helps you know your customers better.

? Join our webinar at 4/15/2022 9AM PST: Learn your customer profile with AI, presented by Panos Moutafis, Co-founder & CEO at Zenus.

Autonomous shopping: 

Zippin’s checkout-free stores are helping basketball and football fans stay out of the crowds (from NVIDIA Blog) 

At Amazon Go store, the shopping cart is built with scale, cameras, and sensors.

Conversational AI

Drive-Thru: Toronto startup HuEx applied conversational AI assistant for drive-thrus. (NVIDIA Blog) 


Fellow Robots, tracking inventory at Lowe’s(NVIDIA Developer channel)

Loss Prevention

Loss prevention through real-time POS theft detection and security. (NVIDIA Webinar)

Back-store automation: 

Automating inventory 

Customers don’t like to see items out of stock, but it is always hard for retailers to decide on inventory before orders come in. Lowe’s applied Veterx AI for inventory management across SKU and store-level forecasts.

Warehouse Automation

Autonomous towing robot capable of towing loads in the warehouse, by KEISUUGIKEN

Warehouse can integrate with human-computer interaction for picking and packaging.

Different AI technologies and tools have been proved to be able to offer internal rates of return higher than historical retail—yet few retailers are moving quickly to implementation. Mckinsey thinks the bottleneck to automation is internally lacking skills and capabilities, not external, at their automation in retail article, 2019.

We are seeing pricing, recommendation systems, large-scale promotions, and category optimization being more widely adopted in e-commerce platforms and it is becoming common sense. However, yet few retailers are moving quickly to AI implementation. Investment in new technologies, integrating new platforms into existing systems, or directly replacing existing traditional architectures, employee learning paths, etc,  perhaps are the reasons we have not seen the retail industry use AI technology generally like e-commerce. Will physical retailers need a revolutionary transformation of the store architecture to get benefit from AI? The answer might be one step further could be better than staying. 

Understand your consumers, driven by data rather than intuition. 

How to collect in-store retail analytics on a massive scale? Analyze customers, promotion activations, and in-store reactions, let’s check out this truly ‘plug and play ’smart camera, powered by NVIDIA SoM and Seeed carrier board for Jetson. Learn more about our partnership with Zenus here. 

Zenus Smart Camera 

  • Measure impressions, demographics, and sentiment 
  • View results on the live dashboard/ use API to build your own application.
  • Never worry about compliance (privacy-by-design) 
  • 17+ million data points from a single camera, cover up to 1,800 ft² (168 m²) 
  • Mount one indoors, outdoors, on-the-go, or in dimly lit environments. Support offline.  
  • One Zenus smart camera analyzes hundreds of faces and processes multiple frames per second. 

Zenus is offering One Month FREE for over three-month subscription for their AI camera.

Interest in Zenus AI camera? Let us know ?here or ?schedule a demo with Zenus!

Seeed is proud to work together with Zenus to offer off-the-shelf, fully integrated solutions. Seeed supports the Zenus AI camera with a high-performance, interface-rich NVIDIA Jetson Nano/Xavier NX/TX2 NX compatible carrier board. The carrier board allows the Zenus smart devices to capture high-quality images for processing by the NVIDIA SoMs via CSI, USB, and other standard interfaces. The combination facilitates advanced facial analysis on edge to deliver valuable insights into consumer behavior.​

Brands currently receive only a snapshot of the in-store customer experience: sales data. The Zenus analytics fill in the missing gaps by offering behavioral insights such as real-time foot traffic, qualified leads, engagement reports, demographics (age/sex), and more. This offers brands a more holistic view of their performance that goes beyond sales data. The facial analytics also tie into conversion data (e.g., POS integration), equipping brands with a complete marketing funnel for in-store and in-person experiences.

Consumer brands and business events can use Zenus plug-and-play analytics to understand people’s behavior in any physical space. Heat maps and sentiment analysis will help retailers know customers better such as analysis in-store promotions.

Heat Maps

Automatically track the total number of impressions (non-unique) and measure average dwell time to understand people’s behavior. Use ethical facial analysis to uncover unique insights about your audience.

Sentiment Analysis

  • Measure the happiness of people over time by analyzing hundreds of faces with over 95% accuracy.
  • Gain valuable insights about which demographics are the most and least engaged.
  • Perform A/B tests to examine the effectiveness of different activations and content to see what works and why.
  • Obtain powerful data and insights for your sales, marketing, and research efforts.

Real Event Analytics by Zenus’ Smart Camera

In the video, Zenus covered most of the functionality provided by their camera analytics tab.

The “Head Count” graph tracks the number of people over time inside the room. The interface also shows the breakdown by sex and age group. You can also obtain a positive sentiment graph. 

Zenus analytics software offers a demographic breakdown over time. The software also displays the positive sentiment level for each demographic group. On the left, you can see the positive sentiment for the two sexes and on the right for each of the age groups.

Latest new arrivals for Edge AI powered by NVIDIA Embedded System

reComputer Jetson series are compact edge computers built with NVIDIA advanced AI embedded systems: Jetson Nano and Xavier NX. With rich extension modules, industrial peripherals, thermal management combined with decades of Seeed’s hardware expertise, reComputer Jetson is ready to help you accelerate and scale the next-gen AI product emerging diverse AI scenarios.

reComputer Jetson is compatible with the entire NVIDIA Jetson software stack, cloud-native workflows, industry-leading AI frameworks, helping deliver seamless AI integration.

? Exciting news! NVIDIA Jetson AGX Orin Developer Kit is available to at Seeed now for pre-order, be the first one get 8x AI performance in hands and experience a giant leap forward for Edge AI.

Seeed Jetson series is compatible with the entire NVIDIA Jetson software stack, industry-leading AI frameworks, development platforms:

Edge Impulse: the leading development platform for machine learning on edge devices, free for developers and trusted by enterprises. Try it now for building models in minutes.

AlwaysAI, provides developers and enterprises a comprehensive platform for building, deploying and managing Computer Vision applications on IoT devices. Get started today with our tutorial, quickly deploy CV applications at the edge.

YOLOv5 ? : You (must have) looked once when you looking for the best object detection application. Check out our latest few-shot object detection use YOLOv5 and Roboflow.

Deci: empowers deep learning developers to accelerate inference on edge or cloud, reach production faster, and maximize hardware potential. Check out How to Deploy and Boost Inference on NVIDIA Jetson.

Nimbus: cloud-based robot development tool

Allxon: Enable Efficient Remote Hardware Management Services for NVIDIA Jetson Platform. Enjoy exclusive 90 days free trial unlock all functions with code H4U-NMW-CPK. Read our partnership story here.

Work with the Seeed Ecosystem

Deploying AI can be faster, flexible, and scalable for everyone. We aim to cover all kinds of AI scenarios at our open-source platform to accelerate industries’ digital transformations. We are looking for partners to join our ecosystem together to deliver solutions to different industries together.

Please contact [email protected] for more partnership ecosystem information.

Learn sth. more

What is sentiment analysis?

Sentiment analysis “refers to the use of natural language processing, text analysis, computational linguistics, and biometrics to systematically identify, extract, quantify, and study affective states and subjective information.”(Wikipedia)

Applying computer vision allows us to analyze crowds of customers and generate happy maps throughout shopping malls. Retails can make use of real-time analytics to drive more precious business decisions such as how to optimize the in-store customer experience, optimized shelves assortments.

Source: Zenus

Amazon Go store

If you have visited the Amazon Go store, you will find there are no human interactions. You just need to walk in, scan the QR code with the Amazon mobile app, carry anything you want to purchase into your basket and finally walk out when you are finished shopping. The store only needs to hire several people to help inventory shelves and answer shoppers’ questions, but there is almost no human interaction. 

The "just walk in, just walk out" technology builds on what Amazon has learned with its smaller Go stores.
The “just walk in, just walk out” technology builds on what Amazon has learned with its smaller Go stores. Source: Lauren Thomas, CNBC
Amazon Dash Cart, source: Amazon

Amazon Dash Cart is similar to the “walk-in and walk-out” process launched by Amazon’s Go store. There is a built-in scale, cameras, and sensors to calculate what you put into the basket. A Dash Cart can hold two bags. When you finish the purchase, Dash Cart will link to your Amazon account to deduct the corresponding payment. There is a screen on the top to display the shopping list and you can even scan the coupon. 

About Author


April 2022