The Dynamic Era of Frictionless Checkout in Retail Stores — Powered by Computer Vision Data Annotation
- Let’s understand why frictionless checkout is making quite a buzz in retail and what are its challenges
- We will take a virtual-readers tour to the companies that have adopted frictionless checkout
- Understanding how data annotation plays the most important role in powering frictionless checkout
Introduction
Physical retail stores have hit hard during pandemics. It’s a fact that they were yet to recover from rapidly changing consumer shopping habits. You must have heard the term “retail apocalypse” that was hitting hard on the ears of retailers. That term defined the state of retail chain owners who faced the closure of business.
But the good news is two out of every three executives surveyed by McKinsey said they have stepped up the adoption of robotics, artificial intelligence, and other powerful technologies in the wake of COVID-19.
With AI-powered use cases like:
- Product recommendations to customers based on their buying behavior
- Inventory and demand forecasting
- Frictionless checkout in retail stores
- Ai-enabled last mile (drone delivery)
- Store footprint optimization
- Optimal planogram efficiency enhancement
And many more. Retailers are now convinced that the adoption of AI-powered technologies can boost their sales like never before.
Research by Google Cloud suggests that for Speciality Retailers the application of AI/ML technologies can potentially drive ~$230–520B in value by 2023.
Let’s narrow our discussion to the most exciting use case in retail. We will discuss the potential of frictionless checkout at retail stores powered by computer vision data annotation. I choose to discuss three exciting topics in this blog:
- Frictionless checkout advantages and challenges for retailers
- Companies who have adopted frictionless checkout at their retail stores
- Data Annotation is the core of Computer Vision Technology that enables frictionless checkout
Frictionless checkout advantages and challenges for retailers
Several years ago, I waited in line for 45 minutes to bill one bottle of soda — An experience that made me never go back to that store again. The main reason behind the long check-out lines is POS (point of sale) systems and devices.
The traditional style of scanning each and every product elongates the waiting time, indefinitely. What we need is a technology like the one demonstrated in the video below:
Standart cognition uses computer vision technology with deep learning and image recognition coupled with smart and accurate data annotation that helps a store track people and products in real-time.
Amazon Go uses the same technology to power its world-famous store and they call it “just walk out” technology.
True frictionless (autonomous) check out should enable shoppers to enter, pick, and leave. For that these are the hurdles a retailer must overcome:
Privacy
You must be aware of the class action law-suite filed against retail giant Walmart a couple of years ago. The claim stated “Wal-Mart’s video recording technology at its self-service checkout kiosks collects “personal identification information” in violation of the California Song-Beverly Act Credit Card Act of 1971 (“Song-Beverly Act”)”.
These are challenges that almost every retailer will face while incorporating new technology. Apprehensions are always there!
Scalability
Adopting frictionless checkout means replacing existing shelves and check out process with sensor-based technology.
Every customer checkpoint needs to be sensor-enabled and that means either the store needs to be closed for the installation time or it has to relocate to a premise with these technologies embedded. Both are commercially not viable.
But an easy way out here is sensor-enabled cameras capturing every detail of the shopping.
Experience
According to many researches done in the US, the two most prominent customer complaints received are:
- Long waiting time in checkout lines and
- Poor customer service
Elimination of long waiting time at checkout counters enables retailers to focus more on enhancing the in-store experience.
In simple words, the retail dynamics are rapidly changing and every retailer needs to embark on a journey to enable unforgettable customer experiences. That’s the only way to increase foot-falls to the store and the correlation is easy — higher footfall is equal to higher sales.
Companies who have adopted frictionless checkout at their retail stores
I think it is some kind of a ritual to start with Amazon Go while talking about frictional checkout. You must agree that they made frictionless checkout to become as cool as it is now!
Amazon Go
January 22, 2018, the historic day when the world experienced a new way to shop for groceries. This day marks the launch of the first Amazon GO store in Seattle. By 2021 there are 27 Amazon Go stores that are located in 4 different cities.
The process at the store is really simple, customers check-in using the scanner in the Amazon app that in turn is linked to their Amazon accounts. Every shelf has a sensor and all the cameras installed around are constantly capturing all the items picked up by customers. In the end, customers can simply leave the store.
Once the automatic checkout process is complete amazon deducts the amount a customer shopped for and shares its receipt over mail.
This video will give you a lot of clarity on how the world’s most amazing retail store functions.
AI Go in Shanghai
AI Go stocks 2000 SKUs raining from fresh produce to packaged goods. The stored is powered by AiFi’s computer vision technology that tracks consumer’s gestures and with data annotation identifies the product in real-time.
AiFi does not capture data via facial recognition. It’s a pure product play and thereby the identity of the shoppers is not revealed. It’s a hybrid approach and here is what Steve Gu, CEO, and Co-Founder of AiFi said in a statement.
“We are delivering a contactless, fast, and enjoyable shopping experience to customers while providing detailed inventory and analytics that make for a very efficient operation for retailers. We are also completely GDPR and CCPA compliant so no customer personal information is disclosed.”
Storelift
The concept of Storelift is way too simple, they intend to fasten the check-out time. They cut all the time out — being stuck in the traffic jam, finding a parking spot, waiting in the check-out line.
They have created a store called BOXY “The Boxy stores are repurposed shipping containers that can be plopped down in various urban neighborhoods that lack good shopping options.”
They believe using AI in this kind of set-up will optimize their costs and increase profitability. “Our vision is really to address all the city areas where there are 10,000 people who have no convenience store,” Storelift CEO and cofounder Tom Hayat said. “We want to be as close to the customers as possible.”
Data Annotation is the core of Computer Vision Technology that enables frictionless checkout
Machine Learning is the fastest growing technology bringing unbelievable changes in the way retailers function. But the most important point to note here is that every machine learning model is built using huge amounts of data.
To create such datasets, an image annotation technique is used. This technique makes objects recognizable to computer vision for machine learning. The volume of data is so high that its practically impossible to label it manually.
Here is an interesting read to know why Manual Labeling is a Daunting Task in Retail and CPG and how Automation Helps
Labellerr provides an ML-based data annotation platform It is a one-stop-shop, high-quality data annotation platform that provides exceptional annotating accuracy to the Retail and CPG industry.
Sign-up here and start labeling by connecting your data from cloud/edge such as GCP, AWS, Azure to integrate computer vision AI into your retail stores or warehouses!
Reach out to us to know more about how you can implement this solution for your business. With our plug-n-play, API solution businesses can customize according to their scenario easily and swiftly.
Originally published at https://blog.labellerr.com on January 21, 2021.
