When retailers are asked about what they want from AI, the answer tends to be quite simple. They want to make their stores smarter, boost efficiency, and increase sales. These were the conclusions of a recent Honeywell report published in August 2023. \nOver 1,000 Retail Directors, IT and Customer Service heads of departments were surveyed across Europe, the USA, Africa, and the Middle East, and nearly 6 out of 10 participants stated that they planned to adopt AI, machine learning, and computer vision in the latter half of 2023.
⁠\n\n
⁠Surveyed retailers agreed that they would be inclined to use new technologies like AI, if they can help them to achieve their goals. The top three reasons for implementing new technology were:
\nImproving Customer Experience (59%)\nDriving Greater Productivity (49%)\nAchieving Cost Efficiencies (44%)\nIt’s all well and good to aspire for better customer experience, staff productivity, and achieving a return on investment, but what does that look like in terms of AI and where would a retailer even start?\nAudience Profiling \nThe first thing a retailer can do is to profile their actual audience and differentiate their shopper (or browser) from the customer (or buyer). AI can now be used to segment shoppers into their respective buying groups, making it easier for retailers to understand their largest and lowest spenders by profile. \nIf a retailer examines their buying metrics, they can then choose to stock products that appeal to the already highest spenders or conversely, they can decide to act on the profiles of browsers that are high in number but not high in sales. \nThey may find that store locations in one city have completely different demographics of shoppers to a store location in another. Yet, if the stores are vending largely the same types of stock, they could be losing out on additional sales from certain key buying profiles. \nStaff Optimisation\nRetailers with a large estate of stores will also have a large number of staff to perform a number of back office and customer facing roles. Many staff will also have dual roles which take them away from the floor. \nKnowing when stores are at their busiest, will allow retailers to assign tasks effectively, in a way that doesn’t leave them short-staffed. It’s fair that retailers may currently make assumptions on the busiest times based on experience, but the ability to see in black and white, which locations are busy at which times, will ultimately save a retail time in staff rostering and task delegation. 
\n\nAd Campaign Effectiveness\nJohn Wanamaker of Macy’s Department store fame is credited with saying “Half the money I spend on advertising is wasted; the trouble is I don’t know which half”.\nThis phrase still resonates in the world of marketing today as there are enormous information gaps when companies launch a localized or national advertising campaign. They will get all sorts of metrics for their online ads, but they have no way of understanding to what extent an advertisement has been successful at attracting their targeted demographics into physical store locations. \nThe most common way to measure ad effectiveness is through sales (ie. Did our sales improve post ad campaign?) But this does not answer any other questions about whether the ad drove product engagement in the stores themselves or attracted new demographics. It also does not shed light on the potential revenue a store could have made on a given day, or how worthwhile the ad was overall to the business.\n\n\nIn-Store Ad Effectiveness\nIn-store advertisements can be overlooked; however, it is one of the most clear and obvious forms of integrating AI in a physical store. What’s the use of having an ad for a product in-store if that product has been sold out for weeks, possibly months. \nThe sparkling wine producer, Rotkäppchen-Mumm, have ingeniously opted for in-store ads that correlate with their shelf availability. The signs content syncs up with sales data to ensure a product that is nearly sold out, does not get any screen time. Oftentimes, the smartest uses of AI can be the simplest. \nThere were also concerns or barriers to the adoption of AI which came up in the Honeywell report, where retailers expressed their main reasons, they would be hesitant to introduce AI, or any new technology for that matter. The main three barriers were highlighted as:\nBudget restrictions (39%)\nDifficulty in demonstrating business value (29%)\nLack of internal expertise to maintain the technology (21%)\nWhat this seems to equate to is a fear of implementation or of stepping away from the status quo. It’s understandable that there would be a fear of misspending company money, being able to show colleagues the value of the AI technology, or being responsible for training in staff in how to utilise the new AI product.\nA solution to this is to ensure that any AI product a retailer is considering comes with things like on-boarding and support included, where the artificial intelligence is catered for or looked after by a partner company, who will deliver that data in context sensitive ways, to specific strata of company management. With a company that acts as a partner, a lot of these barriers or fears for retailers will be quashed.\nIt’s clear from the Honeywell Report that retailers are not necessarily scared of AI as a concept in and of itself, but they are wary of how to get the best out of it. Considering AI has only really come to prominence in the last decade, experienced retailers will want to know that their operation has the budget and means to take on any transformational projects. 

Based on a 2023 Honeywell survey, nearly 6 out of 10 retail directors agreed that AI would be implemented in their retail operations in the very near future.

What Do Retailers Want From AI?

Retailers that are looking to stay one step ahead, are increasingly looking to artificial intelligence (or AI) to help understand the millions of data-points readily available to them in-store. AI has been changing the world around us and retail was one of the first consumer industries to put different forms of it to use. \n In 2019, Harvard, released a white paper, positing that the 2020’s would be the decade of artificial intelligence in the retail space. It’s hard to argue against when the turn of the decade saw the launch of the all-knowing, but still prototypical Amazon Go store. But aside from Amazon, where does the rest of the retail world fit in when it comes to artificial intelligence?\nMost of the talk in retail prior to 2019 seemed to be about the fear of artificial intelligence replacing jobs in the industry, but there was a fundamental misunderstanding over how AI could really be utilized. \nSelf-Scanning checkouts were one of the the first highly visible forms of AI that affected the personal shopper experience. However, checkouts are merely robotic tills, where a customer can scan their own items, and then place them onto a weighing scale before completing their transaction. \nThis is not artificial intelligence as we truly understand it today. AI today is being used to help enhance a store’s operations, to make the lives of floor staff and management easier, and to compliment, not to replace. The most common areas where retailers would like assistance from AI tend to revolve around:\n\nCustomer Experience \n
⁠Customer experience can be enhanced by learning about those in-store. A shopper can be anyone entering a store location, but a customer is the person that actually purchases a retailer’s wares. 
⁠AI can customize a shopper’s experience and retailers can make better decisions for their store if they know who their shopper is, how long they usually browse, and what items tend to be seen together in baskets, or who is most-commonly seen at the till. 
⁠
⁠Having access to this information can allow a retailer to capture more revenue from their highest average spenders, or even increase spending from those shopper groups who are commonly seen entering but perhaps don’t spend as much. 
⁠\nStore Layout
⁠\nStores can be limited by their shop floor and retail space. Some retailers strive to give each store uniformity, and make an experience in one location, as identical as possible to a location halfway around the world, giving their shopper base a sense of familiarity and comfort. Other retailers allow their individual store locations to make theirs their own, whilst selling a mostly identical range of items. 
⁠
⁠Each layout, no matter the brand or location, poses different challenges, and knowing how best to use space to ease a shopper’s journey can be an ongoing project for retail teams. AI can be used to map shop journey, from the way people enter the store and which directions they walk in, to how regularly shoppers make it to certain aisles, changing rooms, or secluded sections of the store, where retailers may like to see more activity or engagement.
⁠\nInventory management
⁠\nAI can play a huge role in inventory management, from knowing precisely how many units of stock are on the shelves to how many units are available to go out on shelves. Retailers have been toying with ways to check shelf availability for years, some larger chains, with the resources to spare, have invested in object detection technologies as a way of knowing when shelves aren’t full or heat-mapping technologies to predict when aisles will get particularly busy. 
⁠
⁠Robotics are in some cases used to scan aisles, resupply shelves, and even assist shoppers in their requests. These technologies have advanced in the previous decade enormously but are perhaps still a little too experimental and expensive for the SME retailer. 
⁠\nForecasting 
⁠\nForbes described it as when data meets AI, and used the newly coined term “omnichannel” to describe changing consumer habits, where a blend of location intelligence, transaction histories, and consumer profiles, all compiled using artificial intelligence was the most effective way to predict consumer trends and spending habits. 
⁠
⁠AI can be used to forecast trends, especially when forms of it are used to show a retailer their shopper demographics and how they interact in-store. Measuring a shopper base while examining sales in tandem, can allow retailers to predict where the next product push needs to happen or where prices should be altered to increase revenues and capture additional spend. \n\nThese are just some of the more notable areas where artificial intelligence is being incorporated into physical retail. While tech leaders have been debating its merits for over a decade, one consensus appears to be that retailers who refrain from embracing AI, can of course continue to function as they always have, but they will constantly have to compete with other businesses in their space who are using AI to their advantage.

Discover valuable insights into the influence of Artificial Intelligence in today's retail landscape, courtesy of VisionR's pioneering in-store solution.

VisionR Blog

What Do Retailers Want From AI?

What are the most common uses of AI in Retail today?