Hello Omni Talk Fans!
This week, I sat down with Gagan Luthra, VP of Product at Impinj, to tackle a problem that has plagued retailers for decades: inventory accuracy.
And we aren’t talking about a small problem. According to IHL, retailers lose an estimated $1.7 trillion annually to inventory distortion. That’s trillion with a T. Between out-of-stocks, overstocks, misplaced merchandise, and inaccurate inventory records, retailers have spent years trying to get a clearer picture of what’s actually happening inside their stores.
But Gagan challenged me to think about the problem a little differently. What if inventory distortion isn’t really an inventory problem at all? What if it’s a data problem? That one distinction ended up shaping our entire conversation, taking us from RFID and Impinj’s Gen2X technology to Zara’s inventory success, physical intelligence, and what all of this means for the future of agentic AI.
Here are my biggest takeaways:
Retail Doesn’t Have an Inventory Problem. It Has a Data Problem.
This was probably my favorite insight from the entire conversation, and after 30 years in retail, I can’t say I’ve heard the problem framed quite this way before. Gagan explained that retailers have spent millions, if not billions, building systems that make decisions based on incomplete inventory information. Just because a shipment manifest says something arrived at a store doesn’t mean it actually did. And just because a system says an item is on the sales floor doesn’t mean it hasn’t been moved, misplaced, or lost.
The problem is that so much of retail’s technology infrastructure still relies on snapshots of what inventory should look like rather than what is physically there. And as retailers increasingly turn to AI to make operational decisions, that distinction becomes even more important. Because AI can only be as good as the data feeding it.
RFID Is Moving Beyond the Inventory Snapshot
We spent some time getting into the technology behind Impinj’s Gen2X, and Gagan did a great job explaining what could otherwise become a pretty technical conversation. At its core, Gen2X improves how RFID tags and readers communicate, making that connection more reliable and efficient. That matters because retail environments are becoming increasingly complex, with more tagged items, smaller tags, and greater demands on reading accuracy.
But the bigger takeaway is what this technology makes possible. Historically, RFID has given retailers snapshots of where merchandise was last seen. Now, improvements in reliability and speed are helping move the industry toward continuous streams of item-level data. Instead of simply knowing that an item was somewhere in the store, retailers can begin understanding where it is, how it’s moving, and what to do next.
Smaller Tags Are Creating Bigger Opportunities
One of the more interesting parts of our discussion was around the physical challenges of RFID. Apparel has been a natural starting point because retailers can attach relatively large tags to clothing. But what happens when you want to track something as small as a lipstick or a bottle of nail polish?
As Gagan explained, smaller tags mean smaller antennas, which makes them harder for readers to detect. And while newer tags are becoming more power efficient, the signals they send back can also become harder to pick up. That’s where advances like Gen2X become important. The better the communication between tags and readers, the more opportunities retailers have to expand item-level tracking into categories that were previously difficult to manage.
Zara Shows What Happens When You Get Inventory Right
Of course, technology sounds great in theory, but I wanted Gagan to give me a real-world example of what all this investment actually delivers. And he pointed to one of retail’s most interesting RFID success stories: Zara and its parent company, Inditex.
By tracking individual items from manufacturing through distribution, stores, and point of sale, Inditex has reportedly improved inventory accuracy from around 65% to as high as 95% to 98%, while reducing inventory counting from days to hours. Gagan also highlighted reported unsold inventory levels below 1%, compared with much higher averages across the broader apparel industry. I’ve seen some of Zara’s RFID capabilities firsthand, including how the technology can track what customers bring into fitting rooms, and it’s fascinating to think about how much operational and merchandising insight that creates.
The Next Frontier Is Physical Intelligence
One thing Gagan made clear is that inventory management is really just the beginning. RFID adoption is expanding beyond apparel into grocery, healthcare, logistics, and supply chain operations, where knowing the exact location and movement of an item can have implications for everything from product authenticity to food safety.
This is where Impinj’s concept of physical intelligence comes into play. The idea is to move beyond simply identifying physical items and toward generating continuous, trusted data that businesses can actually act on. Think about knowing not just whether an item is in a store, but which zone it’s in, where it has moved, and whether it has passed through a particular operational process. That’s a much more powerful foundation for retail decision-making.
Agentic AI Needs Something It Can Trust
We hear plenty about agentic AI and what it might eventually do for retail, but Gagan brought the conversation back to something fundamental: Before AI can make better decisions, it needs better information.
And that’s where things get really interesting. Once retailers have reliable, item-level data, AI can begin moving beyond analytics and into orchestration. Instead of simply identifying a replenishment problem, imagine systems that can coordinate the next steps automatically. Or think about the implications for buy online, pick up in store, loss prevention, and food safety. The opportunity isn’t just to see what’s happening in the physical world. It’s to use that information to make smarter decisions across the entire operation.
Start With the Business Problem, Not the Technology
I closed by asking Gagan what he’s learned about deploying technology across hundreds or thousands of retail locations, because anyone who’s worked in this industry knows that scaling a solution is rarely as simple as it sounds.
His advice came down to two things. First, start with the business problem you’re actually trying to solve rather than finding somewhere to apply a new technology. RFID, computer vision, barcodes, and other technologies can all play complementary roles depending on the outcome you’re after. Second, never underestimate the importance of data quality. As more decisions become automated, accurate and trustworthy information isn’t just nice to have. It’s essential.
And that, ultimately, was my biggest takeaway from this conversation. We spend so much time talking about what AI is going to do for retail, but perhaps the more important question is whether we’ve built the data foundation to let it do any of those things well.
You can listen to or watch my full conversation with Gagan Luthra wherever you get your podcasts.
Apple Podcasts | Soundcloud | Spotify | Amazon Music
Be careful out there,
Chris and the Omni Talk Team
Over 8 years and nearly 200 episodes, the Retail Technology Spotlight Series has featured some of the biggest names and boldest thinkers in retail. Explore the full archive here: https://omnitalk.blog/category/spotlight-series-podcast/
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.