Physical AI.
It sounds like the latest buzzword destined for keynote stages, investor decks, and endless LinkedIn hot takes. Take artificial intelligence, combine it with sensors, connectivity, and the physical movement of products, and suddenly you’ve got a phrase that sounds futuristic enough to capture everyone’s attention.
But here’s the confession most retailers won’t admit: most companies still don’t know where their inventory actually is.
Sure, they know where the system says it is. They know where it was when it arrived at the distribution center. They know when it was scanned at checkout. But what happens in between? The answer is often a collection of assumptions, manual processes, and educated guesses.
And that’s costing retailers far more than they realize.
In our latest episode of Confessions Of Supply Chain Executives, Amir Khoshniyati, Vice President at Wiliot, and Kay Irwin, National Practice Lead for Connected Assets & Smart Supply Chain at AT&T, joined me to discuss why inventory visibility remains one of retail’s biggest unsolved challenges and why Physical AI may finally be the breakthrough that changes the equation.
Here’s what I took away from the conversation.
Retail Has Never Actually Solved Inventory Visibility
The retail industry has spent decades trying to answer what should be a simple question: Where is my inventory?
Retailers have invested billions into barcodes, RFID, warehouse management systems, transportation platforms, and countless other technologies designed to improve visibility. Each advancement has helped incrementally, but none has fully solved the problem. The reality is that most organizations still know where inventory was at specific moments in time rather than where it is continuously throughout its journey.
As Amir explained, traditional technologies create snapshots. They provide information at designated checkpoints throughout the supply chain, but once a product moves beyond those checkpoints, visibility often disappears until the next scan. Inventory effectively goes dark.
The result is what Wiliot describes as supply chain “brownouts,” i.e. periods where products are moving through warehouses, trucks, staging areas, and stores without continuous visibility into their location or condition.
For retailers, that lack of visibility creates more than operational frustration. It creates uncertainty. And uncertainty is expensive.
The Real Cost Of Supply Chain Blind Spots
When most executives think about inventory visibility, they tend to focus on inventory accuracy. The assumption is that if inventory counts improve, operations improve, but this conversation also highlighted something much larger.
The real cost of poor visibility isn’t simply knowing how many units you have. It’s everything that happens when you don’t know where those units are, where they’re going, or what condition they’re in.
For example, a shipment loaded onto the wrong truck can create delays that ripple across multiple facilities. Perishable goods sitting too long on a loading dock can spoil before anyone realizes there’s an issue. Reusable assets such as roll cages, crates, and transport containers frequently disappear into the network, forcing companies to buy replacements while perfectly usable assets remain stranded elsewhere. Even worse, products can be misrouted, delayed, or lost entirely without anyone knowing until the consequences reach the customer.
According to Amir, large organizations can experience losses measured in the hundreds of millions of dollars annually due to spoilage, dwell time, shrink, misrouting, and operational inefficiencies. What’s most striking is that many of these losses stem from the same root cause.
Nobody knew there was a problem until it was already too late to prevent it.
Physical AI Is About Prediction, Not Tracking
This is where the conversation shifts from visibility to intelligence.
At first glance, Physical AI sounds like another version of IoT, RFID, or asset tracking. And while those technologies are certainly part of the equation, Amir made an important distinction: Physical AI is about continuously understanding what’s happening to inventory and using that information to make better decisions.
By combining sensing, connectivity, and AI, organizations can move beyond static snapshots and begin operating in real time. Instead of reacting to problems after they occur, retailers can identify patterns, detect anomalies, and intervene before disruptions impact operations. A shipment headed to the wrong destination can be flagged before it arrives. A temperature-sensitive product can trigger an alert before spoilage occurs.
Retailers already have plenty of data. The value is in turning data into foresight.
Why Connectivity Is The Missing Piece
One of the most valuable insights from the discussion came from Kay Irwin, who emphasized that technology alone doesn’t solve enterprise problems. Retailers often focus on the sensors, tags, and AI capabilities while overlooking the infrastructure required to make those tools work reliably at scale. Collecting data is only half the battle. That data must move securely and consistently across stores, warehouses, distribution centers, and transportation networks.
As Kay explained, connectivity is what makes Physical AI possible. The network serves as the foundation that allows devices, sensors, and platforms to communicate in real time. Without reliable connectivity, there is no continuous visibility. And without continuous visibility, the intelligence generated by AI quickly loses its value.
In many ways, the network is the invisible layer that turns Physical AI from an interesting concept into a scalable business solution.
The Technology Isn’t The Barrier Anymore
One of the most surprising moments in the conversation came when I asked Amir Khoshniyati what he sees as the biggest hurdle to Physical AI adoption.
His answer wasn’t technology. It was organizational alignment.
As Amir explained, the sensing capabilities exist, the connectivity infrastructure exists, and the technology is already proving itself in the market. The bigger challenge is getting the right stakeholders aligned around a common vision and implementation strategy.
Supply chain, IT, operations, and finance all evaluate these investments differently. Success depends not just on deploying the technology, but on creating alignment around the problems being solved, the metrics that matter, and who ultimately owns the initiative.
In many ways, it’s a familiar story. As retail has learned time and time again, technology often moves faster than organizations do.
The 30-Day Action Plan
For executives wondering where to start, both our guests offered the same advice: stop thinking about the technology and start thinking about the problem.
Too often, organizations evaluate new technologies before they’ve clearly defined what they’re trying to solve. But whether the challenge is inventory inaccuracy, asset loss, spoilage, shrink, compliance, or product traceability, the first step is understanding the operational and financial impact. How often does the problem occur? What is it costing the business? And what would solving it be worth?
As Kay noted, the most successful deployments begin with a clear business objective, not a product demo. And as Amir emphasized, many organizations already have infrastructure investments in place that can serve as the foundation for future solutions.
The companies that will be the ones that can clearly identify a costly operational challenge, quantify its impact, and build a roadmap to solve it. The technology is simply the enabler. The business problem is where the real opportunity begins.
The Confession
For years, supply chain visibility has remained one of retail’s most expensive unsolved problems. Despite billions invested in technology, most organizations still operate with incomplete information, relying on snapshots rather than a continuous understanding of where products are, what condition they’re in, and what risks may be developing along the way.
What made this conversation compelling is that Amir Khoshniyati and Kay Irwin laid out a credible path toward changing that reality.
By combining sensing technology, real-time connectivity, and AI-driven intelligence, Physical AI has the potential to create a living, operational view of the supply chain, one that enables organizations to detect issues earlier, respond faster, and make smarter decisions before disruptions become costly problems.
The implications are significant. Better inventory accuracy. Reduced waste. Improved asset utilization. Greater traceability. Stronger compliance. More reliable customer promises.
In short, better business outcomes.
The retailers that embrace this shift will gain more than visibility. They’ll gain a level of operational awareness that has historically been impossible to achieve at scale. And in an industry where margins are tight and execution matters, that advantage could prove transformational.
Watch the full Confessions Of Supply Chain Executives episode with Amir Khoshniyati of Wiliot and Kay Irwin of AT&T to hear the complete discussion on Physical AI, inventory visibility, and the future of connected supply chains.
Available wherever you get your podcasts:
Apple Podcasts | Spotify | Soundcloud | Amazon Music | YouTube
Be careful out there,
– Chris and the Omni Talk team
Music by hooksounds.com
Sponsored Content
Omni Talk® is the retail blog for retailers, written by retailers. Chris Walton founded Omni Talk® in 2017 and have quickly turned it into one of the fastest growing blogs in retail.