The real value of artificial intelligence in-store

“Artificial intelligence in retail” may call to mind in-store robots, automated checkout, and conversational agents that suggest products. But such factors are unlikely to be the first wave of AI in retail. Rather than replacing human workers such as shelf stockers and checkout clerks, initial implications will be about augmenting the retail experience with data and decision-making, TechEmergence research suggests.

This first phase of AI in retail will bring the online offline. Online retail optimizes analytics, marketing, product placement, and product stocking. Websites and apps detect every click, every scroll, and every user action. This information can be fed into data processes that help predict in real time the success of online advertising efforts. For offline retail, AI can help achieve similar goals.

In the next two to four years, AI will create tangible value for retail in three main categories.

  1. AI for stocking and inventory

Inventory optimization is an inescapable consideration for businesses predicated on holding physical items to be sold in-person to customers.

Retailers will be able to use purchase data—both real time and historical, online and offline—to predict inventory needs in real time based on the season, the day of the week, activity at other area stores, e-commerce in that geographic area, etc. This information may prompt daily suggested orders to a purchasing manager, or some AI systems may be approved to make small orders automatically.

Using machine vision for cameras either installed throughout the building or carried by employees, inventory systems will be able to generate accurate, real-time estimates of all products in a store. Such a system could notify store managers of unusual patterns of inventory data, such as suspected theft, or an unusual uptick in sales of a product that may warrant faster reorder.

  1. Behavioral tracking: marketing and product placement

If inventory optimization is about cutting losses, marketing and product placement is about driving revenue. As store environments fill with sensors and systems become adept at tracking valuable information, stores will be able to encourage sales and cart value in new ways:

  • Information from cameras that detect shoppers’ walking patterns and eye movements could be used to analyze interest in products, restructure store layouts, or test new products in locations that get high foot traffic and visual attention.
  • Data on demographic differences between retail locations or differences in days of week or season, as detected by in-store cameras, could be used to adjust product placement. For example, if elderly people shop more on weekdays, products that sell well to them may be placed or promoted more prominently on weekdays.
  • AI systems may become capable of not only detecting patterns in data, but also suggesting marketing actions or promotions. For example, if a product is selling better than competitive brands, a system may suggest more prominent placement or the discontinuation of alternatives. If a product is languishing, a system may suggest a discount and different placement to move it faster and clear space for better-selling products.

In most stores today, cameras are used for little more than security. They require a human to watch a screen in real time in order to take action. In the future, most retail camera systems will be used for computer vision. With limited human interaction, an AI system will detect suspicious behavior and gather broad information about demographics and behavior patterns in-store.

  1. Behavioral tracking: theft

Theft will become increasingly difficult as physical retail locations are “instrumented” to extract patterns. If store aisles are monitored by cameras that detect shopper interaction with an item, then detecting someone putting that product in their shirt or backpack is not a significant step up. Systems may soon be able to detect:

  • Any item hidden or concealed by an individual in the store.
  • Suspicious behavior, as data analysis from previous stores trains the system to recognize behavior patterns common among thieves.
  • Individuals deemed statistically likely to steal based on training data of previous confirmed thieves inside the same store.

The latter two applications are likely to spawn criticism (and maybe rightfully so), but retailers are likely to use them anyway. Security personnel or managers may receive AI system messages prompting them to watch video clips before confronting individuals. At a certain level of consistency and performance, store employees may be confident in approaching an individual in-store simply from a system prompt, referring to footage only as backup.

Systems of this kind will likely employ personal identifiers such as facial recognition, gait analysis, and other factors that can inform law enforcement if the individual escapes the store.

The shift toward efficiency

Offline retailers should prepare to collect and understand more data to improve efficiencies in inventory, placement, and marketing. The biggest retailers will likely be the first to use these technologies. They are most likely to have the budget, the data science-savvy staff, and the ability to collect data at scale from many locations and possibly from large eCommerce sales operations.

As these retailers “instrument” stores to be more like e-commerce sites, retail is likely to make efficiency leaps. Factors other than data will have less impact on retail decisions:

  • Established brand names: A retailer may drop a well-known brand if it knows it sells more of a lesser-known brand and generates the same or lower returns.
  • Sales relationships: A smooth salesperson who actively works a long-term relationship may not be able to win business if the ROI isn’t there for the retailer.
  • Intuition: Purchasing managers, marketers, and store managers who base plans primarily on their “gut” may be at a disadvantage if competitors gradually cut losses and increase sales with nuanced financial decision-making.

While changes won’t happen overnight, expect a general shift toward “instrumenting” stores to make real-time decisions based on hard data. Expect leaders to rely less on “how it’s always been done” and to constantly adjust the retail environment to cut losses and produce gains. Keep a close watch on Amazon and the biggest retailers, particularly those in your market segment. The use-cases that gain traction and deliver value for large players are most likely to define innovative applications for smaller companies.

Daniel Faggella is founder of TechEmergence, an artificial intelligence market research firm. This article is excerpted from techemergence.com/machine-learning-retail-applications.

Faggella’s presentation on how AI will impact retail stores will be joined by other sessions exploring Retail Experience 20/20 at the Shop!X Annual Meeting, Nov. 8-10 in San Antonio, Texas, U.S. For details, visit shopassociation.org/2017-shopx2020.

 
About the Author
Jo Rossman, LEED AP ID+C, is editor of Shop! Retail Environments magazine and the Shop! Retail Environments Insights Center.
in Retail