8 Ways Artificial Intelligence Is Revolutionizing Retail
From analytics to robots, from computer vision cameras to smart shelves, Artificial Intelligence is revolutionising the way retail industry functions. Find out how-
Article by pcmag.com
Artificial intelligence (AI) is making a huge impact on retail as robots are now assisting with inventory checks, monitoring when floors are dirty, and much more. Because supermarkets, for example, often have difficulty keeping up with labor staffing challenges, robots are helping them manage consumer traffic patterns and keep track of price tags. Robots are also delivering business intelligence (BI) on consumer buying behaviors and crowd reaction.
Juniper Research predicts retailers will spend $7.3 billion on AI by 2022, compared with the approximately $2 billion spent in 2018. Both brick-and-mortar and online retailers are now deploying robots. "We are seeing a lot of robotics-based investments come to fruition, especially when you want fast shipping processes," said Pravin Pillai, Global Head of Retail Industry Solutions for Google Cloud.
At "Retail's Big Show," the National Retail Federation (NRF) conference held last month in New York City, Pensa Systems showed off drones that help stores monitor shelf inventory. Nicholas Bertram, President of Giant Food Stores, discussed how the chain will be implementing a Badger Technologies robot named "Marty" in 500 of its stores. AI will combine with predictive analytics to give retailers data about which products will sell the most and how to personalize the products they offer.
Retailers turn to technology to meet customer expectations and deal with tight margins, Google's Pillai noted. Retailers use database platforms such as Google BigQuery and MongoDB Atlas to help them identify top sellers and gain insights on how to replenish inventory. "You could capture information about what products are on the shelf or off the shelf and where people are walking through the store," Pillai said. "They have lots of data they can capture because of the footprint they have, and that leads to being able to build machine learning [ML] models as well."
According to Pillai, product recommendations were a logical first step for AI in retail. "We're seeing the amount of forecasting light up with ML powering it," he said, adding how Japanese fashion clothing brand Uniqlo and other retailers are using Google Cloud to construct forecasting models based on customer demand using ML. Retailers use ML to "better predict which products they should carry," he said. Conversational commerce assistants from companies such as research firm Capgemini combine natural language processing (NLP) with ML models to give guidance to customers during the online shopping process, according to Pillai.
Here are eight more cool technologies that are revolutionizing retail.
Caper Introduces Smart Shopping Cart
Brooklyn, New York-based retail technology vendor Caper has developed a smart, self-checkout shopping cart that uses computer vision, sensor fusion, and three cameras to automatically ring up items placed in the cart. The first time a product is placed in a cart, customers must scan it so that the cart can "learn" the product. After that initial scan, the computer vision features take over and the item can simply be placed in the cart. Caper automatically tallies the price without shoppers downloading an app. When they're done shopping, they can check out by using the credit card readers on the shopping cart. They can use mobile pay or a credit card. Shoppers can then remove their bags and head out.
"We took one of the most conventional tools, which is a shopping cart, and turned it into a 'power shopping cart,'" said Lindon Gao, co-founder and CEO of Caper. "The parts are enabled by sensor fusion and computer vision to directly identify items as they're being tossed into the cart."
Smart shopping carts are a way to incorporate the digital technology into brick-and-mortar locations that has previously been found online, Gao noted. "We want to bring that digital component of online shopping–that visibility and transparency–into physical stores," he said.
The Caper cart will soon also suggest recipes on the built-in tablet for products placed in the cart. Caper is working on that with some large enterprise retailers whose names it can't disclose yet. Caper has so far deployed its smart shopping cart in two stores in New York: all-natural grocer Foodcellar & Co. Market and Gala Fresh Farms. Caper plans to deliver its smart shopping carts to 150 stores in 2019. Caper is currently focused on grocery stores, but plans to expand into other types of retail stores and convenience stores. (Image credit: PCMag)
Spoon Guru Uses AI to Help Shoppers With Food Allergies
Food search and discovery engine Spoon Guru offers a mobile app that uses AI to help allergy sufferers spot the products in a store that contain ingredients compatible with their needs. The app scans digital shelf labels, uses beacons, and integrates with store kiosks. By scanning a bar code, shoppers can learn which foods in a store are safe for them, like whether or not the products are nut-free or gluten-free. The service supports 180 proprietary dietary attributes.
"We combine nutritional domain expertise with AI and ML to make sense of unstructured data," said Markus Stripf, co-founder and co-CEO of Spoon Guru. "We optimize and augment huge amounts of metadata associated with individual products and recipes, and match those against dietary attributes like vegan, gluten-free, low cholesterol, and high fiber." Stripf said he came up with the idea for the company because his wife has several dietary restrictions and was struggling to read food labels in the supermarket.
Spoon Guru is available in Tesco supermarkets in the UK, and the company is in discussions to expand the tool to US supermarkets. "Our platform enables Tesco's customers to accurately and immediately find every product and recipe across their entire assortment that meets their dietary needs," Stripf said. In addition to the scanning capabilities used in brick-and-mortar stores, Spoon Guru also integrates attribute filters on online sites. (Image credit: Spoon Guru/Tesco)
Ocado Uses Google Cloud ML to Handle Customer Complaints
UK-based online grocer Ocado is using machine learning (ML) powered by the Google Cloud Machine Learning Engine to increase the speed of analytics from shopping data and boost customer experience. When customers write to Ocado with complaints, Ocado can use a ML model to sort through and categorize incoming messages, Google Cloud's Pillai said.
Google says Ocado can respond to emails four times faster—a 3.5 percent increase—by using Google Cloud ML technology. Ocado uses Google's TensorFlow open-source software library for ML to gain access to ML algorithms to tag and categorize customer emails. The online grocer can prioritize the emails for response.
In addition to ML and analytics, Ocado uses automated robots (shown above) to help with packaging orders for customers, according to Pillai. The robots use AI tech from Google Cloud. "They have a robotic setup where these systems operate on a grid and they know the presence of other carts," he said.
Ocado is looking to incorporate ML into the robots as part of its automated warehouse to help with recovery from errors and enhance self-testing of the devices. (Image credit: Ocado)
Heasy the Robot Points Customers in the Right Direction
Digital kiosks have existed in places like airports, shopping malls, and train stations for years, but now companies such as Hease Robotics are making them a bit more mobile. The company says mobile kiosks will bring 20 times more interactions than a stationary kiosk. Hease Robotics is producing 20 "Heasy" robots per month, according to Jade Le Maitre, co-founder and Chief Technology Officer (CTO) of Hease Robotics. The company has deployed Heasy the robot in retail locations in countries such as Denmark, France, and Germany. In France, you can find Heasy the robot in the hypermarket E.Leclerc. The company plans to expand Heasy the robot to the US.
Heasy the robot can scan a customer's loyalty card and show deals relevant to that shopper. The company's software then collects data to address a shopping mall's pain points, such as how much time a customer spends in stores. Heasy the robot directs customers in the right direction.
"We have the map of the facility so the robot can give directions to a specific store or special promotions," Le Maitre said. "If it's time for lunch, then it will advise [a] customer to grab some lunch."
The ultimate goal of AI products such as Heasy the robot is to provide the most relevant data to help both customers and retailers, according to Le Maitre. For merchants, that is data on how to sell more products and increase revenue. For consumers, that data is the data retailers gain on what shoppers want. (Image credit: Hease Robotics)
Intel Powers Cashier-Free Stores
Amazon is a leading player in the growing trend of cashier-less retail stores and reportedly plans to open 3,000 new cashier-less grocery locations by 2021. Customers can grab the items for which they're looking and leave stores without going to a checkout counter. In another innovative implementation, Cloud Pick and Intel are collaborating on cashier-less stores in China that incorporate automated door access, cameras, and computer vision to let customers check out without a cashier's help.
The tech also includes weighing sensors, an Intel Core i5 8500T processor, and an OpenVINO toolkit, which incorporates Intel Deep Learning. According to Stacey Shulman, Chief Innovation Officer (CIO) with Intel's Retail Solutions Division, the difference in the type of cashier-less store tech could be in the types of sensors used; some stores might have weight sensors while others might use Bluetooth Low Energy (BLE).
An authentication gate keeps the mobile transaction secure. The AI tech from Cloud Pick is similar to that of self-driving cars, with its combination of computer vision, deep learning, and sensor fusion. In the background, Cloud Pick's Intelligent Retailing Platform (C-IRP) enables retailers to provide data for optimizing their store layout and increasing store traffic. Staff could still be present to walk around the store and assist customers without the need to handle checkout. (Image credit: Intel)
AWM Smart Shelf Pushes Targeted Product Information
Smart shelves are another technology that could keep customers interested in visiting brick-and-mortar stores. One such product, the AWM Smart Shelf, features LED displays and targeted product information. Cameras gather data on shopper behavior and demographics to personalize the videos it displays. AWM can customize the videos according to age, gender, or ethnicity. The AI components keep track of shelf availability within a store. AWM smart shelves incorporate cashier-less checkout using computer vision. The platform senses which products have been removed from the shelves and adds these items to the customers' cart. Shoppers are then charged through their digital wallet. (Image credit: AWM Smart Shelf)
Celect ML Helps Stores Predict Inventory Demands
Lucky Brand is among the retailers turning to ML and advanced analytics to optimize their allocation of merchandise in their stores. Celect's Prediction & Optimization Platform makes this possible with its data modeling and prediction database. Powered by AI technology from the Massachusetts Institute of Technology's Computer Science and Artificial Intelligence Laboratory, the Celect platform helps retailers such as Lucky Brand by pulling data from its customer relationship management (CRM) data and sales transactions.
"Celect is able to predict localized future demand by understanding customer choice, context between products in an assortment, and how demand for a product is affected by the greater assortment around it," said John Andrews, CEO of Celect. (Image credit: Celect)
Zone24x7 Aziro Robot Takes Inventory Counts in Stores
Large department stores are testing a robot called Aziro from Zone24x7. It features an autonomous sensing system that uses radio frequency identification (RFID) to check shelf inventory. Zone24x7 says that RFID can help increase the accuracy of inventory counts and improve the ability to locate items within a store. In addition to a store showroom, the Aziro robot will be used in warehouses and distribution centers.
The Aziro robot features 3D, bumper, and sonar sensors as well as laser range finders to help with navigation around a store. As with many devices these days, the Aziro robot can be controlled remotely. One can do so by using a cloud-based fleet management system, which helps stores organize maintenance tasks.
The Aziro robot runs on the open-source Robot Operating System (ROS) and feeds data into the MySQL open-source database system. It also distributes data by using Apache NiFi and uses the Ubuntu open-source OS. (Image credit: Zone24x7)