In key parts of the value chain, AI, machine learning, and robotics are already being used by retailers. Most importantly, AI technologies have the potential to automate many manual tasks in areas such as promotions, assortments, and supply chain management. Promotions, assortment, and replenishment are the three areas with the most potential in short to medium term. In all of these areas, major retailers are experimenting with AI.
AI is being used by “digital native” e-commerce companies to predict trends, optimize warehousing and logistics, set prices, and personalized promotions. Some companies even try to anticipate customers’ orders to the point of shipping goods before they are confirmed.
The use of AI in retail can have several advantages. To begin with, it enables people to make better decisions by providing more accurate and real-time forecasting. Good forecasts aid in supply chain management, creating effective thematic promotions and optimizing assortment and pricing. Second, AI can improve the efficiency of operations through robotics and process optimization, which increases productivity while lowering manual labor costs. By creating personalized and convenient shopping experiences, AI will enable retailers to increase the number of customers and the average amount they spend.
Will artificial intelligence help traditional, non-digital retailers catch up, or will it widen the gap between the data-driven, agile internet pure players and the lagging historical brands? Retailers’ ability to get on board and secure access to strategic data while reinventing the shopping experience will be critical to success. But, before we look at the conditions for full achievement, let’s look at what the future might be like in 2030.
Retailers know more about what shoppers want
In the future, artificial intelligence could assist retailers in forecasting and automating real-time decision-making. AI can assist businesses in adapting to and mastering an increasingly dynamic market environment by identifying and learning from patterns in large volumes of data from many disparate sources—previous transactions, weather forecasts, social media trends, shopping patterns, online viewing history, and facial expression analysis, seasonal shopping patterns. Machine learning and computer vision can help better anticipate consumer expectations while optimizing and automating supplier negotiations by improving forecasting accuracy.
The impact of AI-assisted forecasting can already be seen. For example, by using an ML algorithm to predict fruit and vegetable sales, a European retailer increased its earnings before interest and taxes (EBIT) by 1 to 2%. The company automatically orders more products based on this forecast to maximize turnover and minimize waste. Similarly, using deep learning to analyze billions of transactions and predict what customers will buy before placing an order, German e-commerce merchant Otto has cut surplus stock by 20% and reduced product returns by more than two million items per year.
When it comes to expanding their physical footprints, AI technologies can also assist retailers in predicting future store performance. Non-digital store sales per square meter are declining as more sales migrate online. If all other factors were equal, retailers in the United Kingdom would need to reduce space by 20% to return to 2010 sales densities. Optimizing storage space and location has become critical for retailers. When deciding where to open a new concept store, a Japanese retailer used machine learning to better understand profitability drivers.
Automation in operations
For AI applications, warehousing and store operations offer many optimization opportunities. Operations automation would be a game-changer for some non-digital retailers, particularly supermarkets. Many supermarkets match online grocers by offering online sales and home delivery. However, they still bear the full cost of physical stores, so the costs of online service—roughly £5 to pick items off the shelves for an £80 order and £8 to deliver them in the United Kingdom—wipe out the industry’s 2% profit margin many times over.
To increase productivity and reduce injuries, autonomous robots can work alongside humans. Since implementing autonomous guided vehicles in its warehouses, Swisslog has reduced stocking time by 30%. Last year, DHL debuted a pair of fully automated trolleys, which follow pickers through the warehouse and relieve them of physical labor.
Machine learning in the store can help optimize merchandising, potentially increasing assortment efficiency by 50%. Using geospatial modeling to determine micro-market attractiveness and statistical modeling to predict and minimize running out of stock, a retailer saw a 4 to 6% increase in sales. These efficiencies would be realized in real-time with machine learning, and they would improve accuracy as they learned from new data.
Ocado, a UK-based online supermarket, is one example of a company that has AI at its core. Thousands of products are guided by machine learning algorithms through a maze of conveyor belts in the retailer’s warehouse and delivered to humans just in time to fill shopping bags. Other robots transport packages to delivery vans, guided by an artificial intelligence application that determines the best route based on weather and traffic conditions.
Many consumers have come to expect personalized, immediate, and spot-on assistance due to the ease, economy, and immediacy with which they have been empowered. Marketers attempting to reach hyperconnected consumers who constantly redefine value by comparing prices online—even and especially when browsing in a non-digital store—will find AI invaluable in the future. With so many people using smartphones, an omnichannel strategy is required, and AI can help optimize, update, and tailor it to each shopper in real-time. Sales could rise by 1 to 5% due to insights-based selling, which includes personalized promotions, optimized assortment, and tailored displays. This type of personalization, when combined with dynamic pricing, can result in a 30 percent increase in sales online.
Thanks to data gathered online, the pure internet players are several steps ahead in targeted marketing. To compete, traditional retailers must begin gaining access to data assets. Electronic beacons have been installed in stores by Carrefour, a global retailer based in France, and Target, a retailer, based in the United States, to collect data on customer behaviors and purchase patterns. They use machine learning algorithms to determine which personalized offers to send to customers while they shop. After deploying beacons in just 28 stores, Carrefour reported a 600 percent increase in in-app users.
AI-enabled personalization could go far beyond targeted promotions as natural language processing improves. In-store virtual assistants could use facial recognition to recognize returning customers, analyze their shopping histories to make recommendations, and communicate in a conversational manner using natural-language processing and generation. Meanwhile, online merchants attempt to give the Internet a more human touch by making personalized recommendations to customers. Stitch Fix, an online personal-shopper service, uses an algorithm to analyze images clients post on Pinterest to better understand their styles, even if they have trouble articulating them online. Online retailers also use smart agents to understand the needs of their customers.
Bringing it all back home
The area of enhanced user experience is probably the most futuristic in terms of AI in retail. Eliminating checkout, deep learning, and computer vision technologies will help store owners compete with online retailers’ one-click convenience. Amazon Go, a Seattle-based test grocery, allows customers to take items off shelves and leave without seeing a cashier or using a self-checkout kiosk. Computer vision recognizes them and links them to products taken from shelves as they enter the store. When customers leave, the system deducts the cost of their purchases from their Amazon accounts and emails them a receipt.
Virtual assistants at home push the boundaries even further. They could alert users when they’re about to run out of a product and suggest that they buy more in the future. Google Home, Google’s smart speaker service, allows customers to place orders with 50 Google Express retailers, including Costco, Whole Foods, and PetSmart. At the same time, Amazon’s Alexa has over 100 third-party service partnerships. Smart home assistant advancements have paved the way for a major shopping disruption. By taking a picture, computer vision assists in identifying desired goods, or the assistant identifies preference patterns from images and videos that consumers like online.
Artificial intelligence technologies could be used to deliver goods minutes after they are purchased on a large scale. The majority of today’s efforts, from companies as large as Amazon to startups as small as Flirtey in Reno, Nevada, are focused on unpiloted aerial drones. In July 2016, Flirtey delivered its first package to a private residence, a box of snacks from a local convenience store. Starship Technologies, an Estonian startup, has taken a different approach in Europe, with six-wheeled delivery robots zipping along city sidewalks at 4 miles per hour. The deployment of drones and robots is dependent on the use of deep learning technology to enable creative problem solving and decision making, as well as airspace regulation, as civil aviation authorities question the use of drones over populated areas and near piloted aircraft’s flight paths.