Perpetual Inventory Intelligence: The Retail Operating Model That's Quietly Replacing the Count

Most retailers still think of RFID as a counting tool. The ones gaining ground have figured out it's something else entirely.

I've spent the better part of two decades working in retail technology — including in IT at lululemon, at Avery Dennison, and as a vice president of supply chain at PVH — and the question I get asked most often by operations and merchandising leaders hasn't changed much: How do we actually know what we have, and where it is, right now?

For most retailers, the honest answer is still, "we don't," at least not in real time. We have a count from Tuesday, a replenishment model built on that count, and a store floor that has been in continuous motion since the count ended. By the time a buying decision is made, a replenishment trigger fires, or a customer asks for a size, the data underneath it is already stale.

This is the fundamental constraint that RFID was supposed to solve. And in many ways it has — handheld RFID programs have meaningfully improved inventory accuracy over barcode cycle counts. But even when it’s well executed, handheld RFID gets you accurate data a few times a week while the store operates every hour of every day. That gap is where retail margins suffer.

A New Category Has Emerged

What I'm watching develop, and what I'd argue is the most consequential shift in retail operations in the past decade, is the move from periodic inventory accuracy to what I think of as perpetual inventory intelligence.

Perpetual inventory intelligence is a new kind of operating model and mind shift. It means your inventory data is always current, not because someone ran a count, but because the store infrastructure never stops reading. Fixed overhead sensors are deployed throughout the sales floor and stockroom and read every tagged item continuously. Because this system doesn’t require scheduling or labor to run, there’s no costly gap in inventory knowledge and accuracy.

While the difference might sound incremental, it’s not. Inventory data that is continuously rather than periodically refreshed unlocks an entirely different set of decisions. Retail teams can detect out-of-stock events in minutes instead of during the next count since handheld RFID errors go uncorrected until the following count. Replenishment triggers fire based on actual floor position, not model estimates. Another benefit is that perpetual inventory intelligence automatically surfaces shrink events (i.e., where items move without a transaction). The power of this kind of system is that it enables retail associates to stop doing the drudgework of counting and start taking actions that create real results.

What 'Perpetual' Actually Means in Practice

The retailers I work with that have made the transition to perpetual inventory intelligence, where sensors read continuously and deliver 99 percent-plus accuracy, describe a similar experience: the first few weeks feel almost disorienting. Store employees say that with perpetual RFID, they just do what their “UFOs in the ceiling” tell them to do, and they trust the system to guide them.

The data doesn't degrade between counts because there are no counts anymore. The system just knows. What they then discover is how many decisions they were making on assumptions they had mistaken for data, and how differently those decisions look when the underlying numbers are actually reliable. Explaining to a new employee how to use this kind of system is very easy as it is, just sensors and a mobile device, with a cloud AI model running in the background. It's like using a map app on your phone. No late night Saturday or Sunday scheduling, no scanning routes, no “paint the fence” or waiting for the beeps to stop, no labor coordination needed.

The Operational Model Shifts Too

This is the part that merchandising and operations leaders often underestimate going in. Perpetual inventory intelligence not only improves your data but it also changes what you ask your store teams to do.

In a handheld model, store associates are a critical part of the data collection process. Their labor creates the inventory record. Contrast that with a perpetual inventory intelligence model, where the infrastructure is what creates the record. Associates shift from generating data to acting on the information. That seemingly minor workflow adjustment is really a meaningful change in how stores are staffed, managed and measured.

The retailers moving fastest on this transition are the ones treating it as an operating model change instead of as a technology deployment. They're redesigning exception workflows, retraining floor leads, and rebuilding replenishment logic around the assumption that inventory position is always known — because it is.

Where This is Going

In my previous writing for Total Retail, I made the case that inventory accuracy is the foundational layer for agentic AI in retail and that AI agents making autonomous decisions about replenishment, fulfillment, and allocation can only be as good as the data they're acting on. I stand by that view. However, I'd go one step further now: more than just AI-readiness infrastructure, perpetual inventory intelligence is the operating baseline that modern retail requires, with or without AI.

The retailers that will define the next decade of performance are the ones that abandoned traditionally time-intensive handheld RFID counting altogether and built stores that simply know what they have, every minute of every day.

That shift is already underway. The question for every operations and merchandising leader reading this is whether their organization is leading the change or watching it happen in the store next door.

Jonathan Aitken is senior vice president, RFID Center of Excellence at RADAR, where he leads industry-wide RFID best practices and helps retailers scale real-time inventory across their store networks.