HayDmitriy/Depositphotos.com
November 7, 2025
Which Tech Will Solve In-Store Inventory Woes?
Roaming robots, on-shelf cameras, and sophisticated handhelds appear to be battling over how to solve understocks at the store level.
The most controversial is the use of robots. The 2025 SPAR Consumer Survey found 71% of shoppers uncomfortable or unsure about the idea of roaming robots with cameras in stores, with women expressing more discomfort than men.
In 2020, Walmart, one of the biggest adopters of this technology, announced it was ending a contract that saw shelf-scanning robots appear in some 500 of its stores. The Wall Street Journal reported at the time that as more people began shopping online, Walmart had “more workers walking the aisles frequently to collect online orders” that could also perform inventory checks. Walmart also reportedly had concerns about what customers would think of seeing robots in stores.
Yet Shop & Stop & Shop and Giant, both owned by Ahold Delhaize, have had robots patrolling aisles to check for spills and stock for several years. Kroger also began piloting robots at some stores earlier this year, with BJ’s Wholesale Club, SpartanNash, and Schnuck Markets also among those employing inventory-checking robots.
New research from IHL Group, in partnership with Brain Corp, a robotics firm, found robotics ranking as the top solution to improve inventory accuracy, ahead of handheld RFID, mobile image capture, fixed cameras, or barcode scanners. The survey of about 200 retail executives found 72% of retailers are ready to deploy in-store robots.
Robots May Not Be the Only Answer to Inventory Struggles in Retail
However, Walmart Canada, Asda and Morrisons in the U.K., Metropolitan Market in Washington, and Australia’s Woolworths are among retailers that have rolled our or begun trialing shelf cameras using computer vision to monitor stock levels in real time. At least one Woolworths’ store installed 500 mini cameras on product shelves.
In an ongoing trial at five Asda stores, the cameras are mounted on poles on top shelves, each monitoring around eight feet of shelf space opposite. Kevin Johnson — CEO of Focal Systems, the technology provider — told The Grocer earlier this year that “deep learning breakthroughs in computer vision have enabled the precise identification of products, shelf conditions, and anomalies in highly complex and dynamic retail environments.”
News of Asda’s test has drawn privacy concerns, including that they might be used to monitor shoplifters.
Meanwhile, handheld scanners have long been used by store staff to not only more quickly read SKU or UPC codes versus counting by hand with inventory count sheets, but also quickly get the information to retail POS systems. With AI and other advances, the handhelds are seeing an upgrade.
In September, Starbucks announced it was rolling out AI-powered automated counting via scanning tablets across its North American fleet. Combining computer vision, 3D spatial intelligence, and augmented reality, the technology has proven an ability to count inventory eight times more frequently for better real-time visibility.
Deb Hall Lefevre, Starbucks chief technology officer, said in a blog entry, “In just minutes, it identifies what’s available, flags low-stock items and, soon, will even automate restock orders. It’s a farewell to manual tallies — and a hello to smarter, more seamless operations.”
Discussion Questions
Which technology solution has the most potential to reduce understocks and support real-time inventory visibility?
Are robots, shelf cameras or scanners the answer?
Poll
Loading …BrainTrust
Tom Ryan
Managing Editor, RetailWire
Recent Discussions
More Discussions
Will Ford’s New Retail Experience Deliver Results?
Ford is revamping its global dealership model, in terms of physical space and other aspects. Will…Will Target’s Rebalancing Act Solve its ‘Sloppier’ Stores Problem?
Target is making an operational move to take fulfillment duties away from certain stores. Will this…Have Retail Loyalty Programs Moved Beyond Discounts?
Loyalty programs are evolving. Are discounts still the top motivator for shoppers?
The right solution depends on identifying the underlying problem. If inventory gaps are caused by a lack of monitoring on the shop floor, then shelf-scanning robots or video surveillance combined with AI can help. If gaps stem from improper forecasting, then AI-powered prediction can assist. But if stock shortages result from misaligned buying incentives (here’s looking at you, Target), none of these solutions will help until those policies are changed.
“News of Asda’s test has drawn privacy concerns, including that they might be used to monitor shoplifters…”
Heaven forfend that we should impinge on the privacy of, ahem, shoplifters!
This is a strange question (yes, I still exist). Inventory problems will be solved by a combination of people, processes and technology. And maybe robots. But it’s a deeper question than can be answered here and is deoendent on segment, culture and store footprint.
and so my answer is (since emojis don’t seem to not be accepted) a shrug of the shoulders with a bit of “ I dunno.”
Not only is there no right answer, but solutions vary by segment. The issue is massive, though, and retailers who figure it out, like apparel companies using item-level RFID, will reap the benefits.
Robots are identifying problems that shouldn’t have happened in the first place, and unless the inventory is in an in store back room, not much can be done to fix the problem in a reasonable time frame anyways. Useful – sure it’s a feedback loop. Problem solving technology – not so much.
When it comes to maintaining on-shelf availablity and planogram compliance, “detect-and-remediate” tactics will NEVER outperform a sound computer-aided ordering discipline.
Whatever sensing technology is deployed – fixed cameras, roving robot cameras, hand-held scanners, RFID, or a combination – the data must flow into a store level demand forecasting system that is tuned to minimize inventory inaccuracy in the first place.
Computer vision is getting better and smarter thanks to AI, so I’d bet that cameras will continue to gain in utility for this purpose.
Taking into account impact on customers as the most important priority in operations, roaming robot scanning must be done after hours, not during prime hours to avoid increased shopper frustration. Grocery shoppers’ time and convenience must be protected when they only have 10, 15, or 20 minutes or less. This is where cameras and AI are a little more seamless.
Robots and cameras viewing shelves in store can do more than just check inventory – they can give a perspective on planogram compliance and potentially also get a perspective on when during the opening day items are most at risk from being not in shelf.
The challenge with this technology – as is always the case in retail – is the multiplier – roll this out across hundreds of stores and the cost becomes eyewatering – meaning the business case must be beyond watertight.
It is technology that has been discussed and available in multiple forms for over 10 years – shelf edge cameras, moving robots, associated held cameras – all of these options have existed.
Often a major barrier is the store technical infrastructure and the ability to leverage the real-time – or near real-time – data usefully downstream to make actual useful decisions that are immediately actionable. But then which department pays for it? Retail operations because it helps them move product from store-room to shelf? Supply chain because it helps them serve the retail operations team better? Merchandising because they can better understand how their promotional activity is working?
And this is part of the challenge – a little like a paying for a bridge in a community – no-one wants to pay for it – but everyone wants to benefit from it by travelling across it.
Inventory problems do not come with a one-size-fits-all solution. So, an unsatisfying “it depends” is the answer. Leaving aside the classical supply/demand and push/pull approaches, or supply chain disruptions, we have the search for technological silver bullets, which don’t exist. Robots can accurately scan shelves, but they are expensive and may be off-putting to customers. Shelf cameras can quickly identify holes, but they raise consumer privacy issues. Handheld devices must rely on labor and schedule compliance. Whatever a retailer chooses, it must fit and operate as a cohesive system.
This debate isn’t really about cool technology; it’s about solving the problem of inventory, where stock is gone but the computer hasn’t registered a sale. Current replenishment systems are useless against this, so the decision must come down to Return on Investment (ROI). Big box stores are the only ones who can justify the massive cost of roaming robots, as their sales volume makes the technology worthwhile. Most other businesses should look to simpler solutions, like AI-powered handhelds, to quickly prove a sales-linked return. Remember, technology is just needed for inventory check, but staff service and empathy are absolutely irreplaceable.