Laurent Giraid is a distinguished technologist whose work at the intersection of machine learning and physical environments is redefining how we interact with the world around us. As a specialist in Artificial Intelligence and the ethical implementation of emerging tech, Giraid has closely monitored the evolution of systems that bridge the gap between digital data and physical reality. In this conversation, he explores the massive economic drain within the retail sector—a 15-billion-dollar problem in the United States alone—where human potential is often buried under the weight of manual inventory tracking. Through the innovative application of radio frequency identification and cloud-based algorithms born at MIT, Giraid explains how a new category of “Spatial AI” is transforming global fashion giants, allowing employees to spend less time in the stockroom and more time engaging with people.
The themes of this discussion focus on the transition of AI from digital processing to physical perception and the tangible impact of indoor localization technology. We delve into the origins of these systems within academic research labs, the logistical hurdles of scaling to hundreds of international locations, and the future potential for this technology to revolutionize manufacturing and robotics.
Retail workers often spend half their shifts managing inventory, costing billions annually. How does this technology transform that burden into a streamlined process for both the employee and the customer?
The reality of modern retail is far more labor-intensive than most shoppers realize, with nearly 50 percent of all working hours currently dedicated to the grueling task of managing inventory. In the United States, this inefficiency translates to a staggering 15-billion-dollar problem every single year, as employees are forced to spend their time combing through crowded stockrooms and sprawling shop floors just to find a single item. We have all experienced that frustrating moment where a store associate disappears for 20 minutes to check the back for a specific shirt size, only to return empty-handed. Our technology changes that dynamic by using wireless signals from radio frequency identification tags to pinpoint the precise location of every item in real-time. By turning those invisible signals into a searchable map, we allow the associate to find what they need in seconds rather than minutes, which directly improves the customer experience and allows staff to focus on higher-value interactions. This isn’t just about saving time; it is about reclaiming the human element of retail by removing the invisible wall that disorganized inventory creates between a brand and its patrons.
The journey from an MIT research lab to a global deployment involves a significant shift in perspective. Can you walk us through how the focus moved from fundamental wireless research to a practical solution for the retail world?
The core of this innovation actually dates back more than 15 years, rooted in a deep academic pursuit of how wireless signals can be used to sense the world in ways that were previously thought impossible. While at the MIT Media Lab, the research team, including individuals like Fadel Adib and Isaac Perper, spent years developing the fundamental machine-learning algorithms that could translate raw RFID data into meaningful location patterns. The real turning point occurred in 2021 during the National Science Foundation’s I-Corps program, where the team conducted extensive interviews with potential customers to identify the most painful problems they faced. It was through those conversations that they realized just how broken inventory management was—not just as a technical glitch, but as a massive operational bottleneck. After founding the company in early 2023 with support from a small business award and MIT’s Venture Mentoring Service, the focus shifted toward making the technology scalable and cost-effective. We had to move away from the complex, fragile setups of a laboratory and create a robust product that could thrive in the high-traffic environment of a global fashion retailer.
Many retailers are hesitant to adopt new tech because of the hardware costs and installation headaches involved. How were you able to leverage existing handheld readers to create a ‘Spatial AI’ that maps a store’s inventory in real-time?
One of the most critical decisions made early on was the “big bet” that we could build this entire intelligence layer on top of hardware that retailers already own and use. Most large stores already utilize handheld RFID readers to check what is in or out of stock, but they were missing the spatial context—knowing where those items were actually sitting. Our platform integrates directly into the existing inventory applications or custom employee apps, acting as a sophisticated brain that processes the data generated during routine scans. All the heavy lifting and sophisticated location algorithms sit in the cloud, which means the store doesn’t need to install expensive new sensors or intrusive infrastructure. This “Spatial AI” allows the system to perceive the environment, deciphering the indoor location through wireless signals and generating a live map of the store. By keeping the interface simple and the hardware requirements non-existent, we removed the traditional barriers to entry that usually stop innovative tech from reaching the shop floor.
Scaling a technology to over 700 stores across 15 countries is a massive logistical feat. What were the biggest challenges in making this system adaptable to diverse retail environments like ZARA or Pull&Bear?
The beauty of our current deployment, particularly with a global powerhouse like Inditex and its brands like ZARA and Oysho, is that the system is designed for rapid, friction-free expansion. When you are dealing with hundreds of stores across 15 different countries, you cannot afford a deployment process that requires a team to travel to every location or spend days on setup. We have reached a point where we can add a new store to the platform in about one minute, essentially by flipping a digital switch and preparing the data for the customer. The challenge wasn’t just the speed, but the diversity of how different stores organize their inventory, from specific bins to categorized shelves. Because our machine-learning models are trained to work with any type of RFID and any store layout, the system adapts to the environment rather than forcing the retailer to change their habits. Since our first major contract in 2025, we have proven that the cloud-based model allows us to scale at a pace that keeps up with the fast-moving fashion industry.
While fashion and retail are the current focus, the potential for indoor localization seems limitless. What is your forecast for the future of this technology in other sectors?
I believe we are entering a phase where AI will no longer be confined to the digital screens of our computers and phones, but will fully inhabit the physical world. While we are currently focused on reaching tens of thousands of retail stores over the next year, the underlying technology foundation we have built for “Spatial AI” is incredibly versatile. Our models are already capable of working with Wi-Fi and Bluetooth signals, which opens the door for massive improvements in manufacturing, logistics, and even the way robots navigate complex warehouses. Imagine a robotics company where every machine has a perfect, real-time understanding of every component’s location, or a logistics operator who can track a package’s precise movement through a massive hub without manual intervention. My forecast is that within the next decade, spatial intelligence will become a standard utility, as fundamental to the physical economy as GPS is to the digital one. We are moving toward a future where machines can perceive their environment with such clarity that the traditional friction of searching for “things” simply ceases to exist.
