Over the past 12 to 18 months, no retailer has been able to send senior executives to talks without first preparing a partisan line on AI, especially generative AI. As in every other field, as the AI hype cycle continues to expand, the technology's “magic bullet” potential has been played out to varying degrees.
The food sector is no exception. At the Morgan Stanley Technology, Media and Telecom Conference last month, Shreshku Kumar, Walmart's global CTO and chief digital officer, spoke passionately:
We're particularly excited about using the latest generation of AI to power search and drive a more conversational experience with our customers. So part of adaptive retail is first understanding what your customers want so you can adapt to it. This is where something like Gen AI truly comes into play. This allows for natural interaction and understanding.
Kumar added:
Gen AI is already in use. gen announced AI-based search. We use large-scale language models trained on world knowledge. We built it and trained it in-house on our proprietary platform to power key parts of the searches that occur on our app. So it's already a reality.
Amazon
That said, recent events may be interpreted as indicating that more realism is creeping into the assumptions of AI's potential, at least for now. Back in 2016, Amazon was talking about a revolution:
What if cutting-edge machine learning, computer vision, and AI were woven into the very fabric of your store, eliminating the need for lines?
This coincided with the launch of the first Amazon Go, “a new kind of store with the world's most advanced shopping technology.” No lines or checkouts required. Just pick it up and go. ”
This has evolved into widespread use of Amazon's Just Walk Out technology in 47 of its 64 Amazon Fresh stores. The experiment came to an abrupt end earlier this month when The Information broke the news that the technology would be phased out in the United States in favor of self-checkout shopping carts. (It will continue to be used in some Amazon Go convenience stores in the US and smaller His Fresh stores in the UK).
The news follows earlier reports that the free checkout strategy relies on a network of cameras operated by 1,000 outsourced workers in India. Amazon has acknowledged that it has Indian workers who annotate video images to help advance machine learning models.
Others have been here before too. Back in 2019, British grocery chain Sainsbury's discontinued its checkout free trial at one of its London stores, a move that the company's chief digital officer said was “a big deal for us. It was positioned as an experiment at the time of its launch, rather than a new format. However, the conclusions drawn from that experiment were as follows.
It is clear that not all customers are completely ready for open-ended stores.
but…
Of course, that was before the generative AI hype bomb exploded. Tesco, the UK's largest grocery store operator, has opened its first GetGo free checkout store in 2021. Back in 2024, CEO Ken Murphy made it clear that AI technology would play a key role in the company's future strategy.
We are continually looking for new and improved ways to better serve our customers. We recently introduced new AI-based tools to improve our retail proposition. This allows for more tailored products depending on store location and demographics. This means that stores will be stocked with more and more products that customers want to buy.
We are also now using data-driven software to more efficiently drive transportation scheduling and inventory assembly processes in our supply chain. We expect AI-driven solutions like this to grow as part of the way we do business in the future, supporting colleagues to focus on the most value-added parts of their roles.
The focus here is on operational efficiency and backend optimization, perhaps a more practical proof point for the role of AI in the grocery sector. Back to Walmart's Mr. Kumar.
It's not just about artificial intelligence itself. The real lock comes when we use breakthroughs like Gen AI and large language models in conjunction with a host of other technologies. Here is an example of how to forecast demand. We've been building our own ML models, deep learning models, that help us understand what customer demand is going to be, short-term, long-term, regionally, and actually even at the household level. So, [these are] Traditional model. By injecting this with the latest breakthroughs from Gen AI, we can really understand it not just at the product level, but at the basket level and the household level. When you put these pieces together, a lot of magic starts to happen from there.
Other companies are taking similar approaches with their own AI initiatives. In other words, it's practical, practical, and not some fancy “up front” thing. Britain's Morrisons has installed thousands of AI-enabled cameras from US AI company Focal Systems across its store network to help store staff stock shelves faster. The camera scans the aisles and flags inventory using a series of categories such as out of stock, out of stock, etc. Non-compliance with planogram (i.e. something is in the wrong place). Low stock. And it's back in stock.
Some companies are turning to AI to tackle the very real problem of self-checkout theft. A UK co-op believes it lost £70m to retail theft last year alone and is now deploying AI-powered CCTV to track what shoppers put in their bags at the checkout. It has been introduced. In the US, Target is facing a significant increase in theft, so it's no surprise this week that Target was reported to be introducing his AI camera technology to its self-checkout registers.
my view
This operational use of AI technology makes pragmatic sense in the retail industry, and despite the somewhat gleeful media coverage of Amazon Fresh's decision, the disruption AI technology creates Regardless, the benefits will be realized in the short term. AI generation hype. Lastly, I would like to leave a comment from Mr. Kumar of Walmart.
The way I see this going is that next-generation AI will just be another tool, just another technology. It will be hidden and available everywhere. But that alone is not noticeable. Everyone uses Excel now, just as other technologies have become practically invisible. So I think generational AI will just become part of the toolbox and people will just adapt to it.
