How Shopic’s Computer Vision AI Reduces Friction in Grocery Shopping
Thanks to advancement in computer vision and AI, the next generation of in-store experiences will blur the line between e-commerce and physical stores
Digital transformation has impacted several business sectors, and the food retail sector is no different. From cashier-less checkouts to virtual storefronts, digital solutions have revolutionized how people shop for groceries.
The humble shopping cart is the latest asset to undergo digitalization. Amazon and Kroger have tried smart shopping carts that scan and checkout customer purchases. However, these carts tend to be expensive and pose storage challenges. One Israeli startup believes it has a solution.
“We want to close the gap between e-commerce and physical shopping,” says Eran Kravitz, CTO and co-founder of Tel Aviv-based Shopic. “Supermarkets have specific needs, and we felt that utilizing computer vision and AI can help grocers provide customers the experience they are after. This is how Shopic’s unique smart cart solution was born.”
Shopic’s solution solves key smart cart issues for supermarket chains. Here’s how the startup manages to do it.
Ringing it up with computer vision
The first generation of self-checkout cart solutions required significant effort from customers. For instance, once a customer pulled an item from shelves, they would have to scan the item correctly.
This process posed significant hurdles.
Customers would need assistance if they scanned an item twice or if the barcodes were damaged.In addition, they might forget to scan items or scan them incorrectly, leading to billing issues.
Other tech solutions have tried mitigating these issues by using cameras to capture items in shopping carts and automatically scanning them. However, given the lack of sophistication in many AI engines, items placed at odd angles, while the cart is in motion, or in large numbers would slip through, creating losses for retailers.
Shopic’s solution clips onto regular shopping carts and uses two front-facing cameras connected to an advanced AI system. Retailers train the AI during the onboarding process by placing items in front of the cameras a few times. Once finished, Shopic’s computer vision accurately identifies over 98% of the items that it has been trained on, and it detects when an item is added or removed from shoppers’ carts with over 99% accuracy, according to the company’s website.
Shopic offers additional support by maintaining a 50,000-strong product database, adding 10,000 items weekly. Best of all, supermarket chains can sync their item lists and updates across all locations.
Kravitz points out that Shopic’s AI continuously learns as customers use it. “Our AI uses computer vision to differentiate between ten of thousands of products,” he says. “The system automatically detects packaging design changes and other variables so it can recognize new items within a day or even a few hours of it going on shelves. Our database is, therefore, always up to date.”
Making analog smart
Smart carts are a great solution on paper but can be costly to implement. For instance, Amazon’s Dash Cart units are manufactured from scratch and need significant upfront investment. Grocery branches implementing these types of carts need to designate indoor storage and charging space for them, while also ensuring customers use them appropriately.
Customers typically leave shopping carts unattended in parking lots. Inclement weather, accidents, and damage often occur. A smart cart with sensitive tech can hardly withstand such conditions. As a result, making the leap to installing smart carts is impractical for many supermarket chains.
Shopic solves this problem elegantly by clipping its device onto regular shopping carts. The company’s devices need a charging station, a small change for supermarkets to incorporate given the ROI from interactive shopping experiences. Customers pull the devices from the wall and clip them onto their carts, instantly receiving a digital shopping experience in a physical store.
From the supermarket’s perspective, incorporating a digital shopping experience is now cost-effective. More importantly, customers experience minimal disruption. Shopic’s device features a built-in scan-and-pay feature, helping customers skip lengthy checkout lines.
Thus, supermarkets can reduce the burden on their cashiers, offering their customers the ability to self-serve. The combination of offering seamless digital experiences while minimizing disruption for customers and supermarkets has generated plenty of attention for Shopic. The startup recently raised $35 million in a Series B funding round, led by Qualcomm Ventures and other notable participants.
Adding value through data analysis
Shopic’s AI does more than identifying items placed in shopping carts and processing transactions. The system also maps stores, paving the way for a customized shopping experience. For instance, supermarkets can create shopping guides based on customer-uploaded lists or present product offers, whether in-house or third party. Shopic thereby turns the humble shopping cart into a revenue stream for supermarkets.
The system can also welcome repeat shoppers by syncing previous shopping lists and suggesting alternatives or creating a store map that minimizes time spent searching for products.
Supermarkets can leverage other use cases too. For instance, they can analyze shopper data and offer up-sells, boosting revenue. They can capture buying trends quickly and modify their supply chain and staffing schedules to account for changes.
“Thanks to advancement in computer vision and AI, the next generation of in-store experiences will blur the line between e-commerce and physical stores,” Kravitz says. “The data that our cameras and sensors gather, and that we then later analyze, will help us open new worlds for shoppers and retailers.”
Ensuring security while enhancing experiences
Devices that capture shopping data aren’t a novel concept. However, these devices are notorious for overzealous data sharing. Kravitz is aware of the issue and insists that Shopic does not use data for monetization beyond the platform’s boundaries.
The company charges retailers a subscription fee for its hardware and software, ensuring customer data never leaves its systems. Shopic’s devices are currently used in Israeli supermarket chain Shufersal, and additional pilot programs are underway in regional chains based in North and South America.
“Our aim is to transform the in-store shopping experience into a fun and easy experience, much like how we purchase products online,” Kravitz explains. “The future will see ecommerce and physical stores merge into an all-encompassing, AI-driven experience, and Shopic will be a part of that change.”