Sensors and vision enhance store performance

RFID-adjacent IoT and computer vision are converging on the same 2026 agenda for retailers--powered by real-time, machine-readable truth, writes Miya Knights, Retail Technology Publisher

Strip away the "agentic" buzzwords, and the annual Retail's Big Show, held by the US National Retail Federation in New York last month still revolved around a familiar retail problem: stores are being asked to do more, with less—more fulfilment, more self-service, and more margin pressure, while meeting ever-higher customer service expectations.

Two technology tracks at the show stood out as genuinely operational: item- and asset-intelligence, such as radio-frequency identification (RFID) tags and Internet of Things (IoT) sensors, and store-wide computer vision. In different ways, both are turning stores into environments that can be measured, managed and protected in real time.

RFID and IoT move from visibility to execution

RFID has already proved its value for inventory accuracy in apparel and general merchandise. Still, the NRF message from IoT platform and RFID provider Wiliot was that retailers increasingly want richer signals than "seen/not seen," especially in grocery, perishables and high-churn operational environments where RFID is harder to deploy or less effective.

In a post-show insights call with Retail Technology magazine, Amir Khoshniyati, vice president of marketing for Wiliot, argued that his company does not offer point solutions but a data platform that can ingest and operationalise signals from the physical world: "Wiliot is a platform company… [built] to ingest data from all different aspects of a retailer's operations."

Wiliot used NRF to introduce its Gen3 IoT Pixel, a battery-free sensing tag designed to expand the types of store and supply chain signals retailers can capture. Khoshniyati said the Gen3 tag expands beyond location and temperature: "It extends beyond temperature now to humidity and light. You can triangulate a lot of the different data points within that, and within the platform, you can then figure out exactly, proximity-wise, where the motion is behind."

This capability matters to store productivity because it shifts the conversation from periodic audits to continuous operational visibility. For example, if retailers can reliably detect dwell time, route patterns, and environmental conditions, they can reduce exception handling, such as missed cold-chain compliance, late replenishment, mis-routed assets, wasted colleague time searching, and the "phantom stock" that undermines online fulfilment.

Khoshniyati also positioned the company's technology as complementary to RFID economics and performance expectations, making direct claims about range and cost: "The Pixel delivers roughly 30 to 40-metre read ranges, which is about two and a half to three times what RFID does," alongside "the reduced cost structure, so it's about half the price, targeting 10 cents at volume." He added: "It has transitioned to an aluminium material stack up versus copper, so we now have a sustainable aspect to the tag as well."

For grocery and convenience operators, he was explicit about where Wiliot believes the near-term win sits: "Right now we're playing very well in perishables. So food, grocery, quick service restaurants… are a very nice sweet spot, because that's where RFID has limitations, and the value can be really brought forward quickly."

"Source of truth" integration

A recurring NRF theme for store tech is that no single platform can become the system of record for everything; what matters is how quickly data flows into operational tools that colleagues already use.

Khoshniyati leaned into this interoperability requirement directly: "You can use the Wiliot platform as your one source of truth, if you choose… but then you can relay that information via API to any third-party platform." He continued: "We will decrypt that information, we will synthesise it, and then we can pass it over to any kind of end destination, relatively simply."

Wiliot aims to deliver on what store leaders are trying to achieve: fewer wasted steps. If signals from RFID and IoT can reliably trigger tasks—replenishment, recalls, rotation, pick-path optimisation, exception resolution—then productivity gains become measurable and repeatable.

Khoshniyati also pointed to the next step in development: predictive operations, built on accumulated high-quality data rather than isolated events. "I think we're not far away from it… If you have enough quality data, you can then start to make the right predictive assumptions, and then that analysis will be very helpful."

Turning cameras into control systems

If IoT and RFID aim to reduce friction by improving truth, retailers are increasingly deploying computer vision in their stores to protect truth, specifically, by stopping losses that occur when store journeys become more self-serve.

Retail computer vision and checkout-free point-of-sale provider Trigo Retail's NRF focus was on store-wide loss prevention, using existing ceiling cameras to detect "what happened" in real time rather than "what was scanned".

In its NRF announcement, Trigo describes a system that "compares each scanned item with what the shopper picks up," triggering alerts when an item is taken but not scanned at self-checkout, with the stated goal of "redefining shrinkage control in-store" while maintaining a frictionless experience.

In a separate interview, Daniel Gabay, CEO and co-founder of Trigo, summarised why the company believes this technology is now scalable: "Cameras were installed in stores decades ago… but no one actually practically used the camera to do things." He added: "There are the eyes, but there is no brain, and we add this brain layer in the store to enable retailers to use them proactively."

Crucially for food retailers, where trust and compliance expectations are high, Gabay was explicit about what Trigo's approach is not: "Trigo is in any way not even close to doing [face recognition]. It is fully, fully anonymised." His comments align with Trigo's published claim that the solution "never uses, collects, or stores any biometric data."

He also argued that the system's intent is to reduce false positives by focusing on item-level truth rather than subjective interpretations of behaviour: "Trying to take measures by which you look at suspicious behaviour can generate a lot of false positives."

Productivity, shrink and CX convergence

What links RFID/IoT sensing and computer-vision loss prevention is that both are being positioned as levers that improve the unit economics of store operations while protecting the customer experience. Inventory truth reduces wasted labour and prevents disappointing "out of stock" moments.

At NRF, Wiliot's message suggested that the future store is one where "physical AI" makes routine tasks the exception. Trigo implied that the future store is one in which exceptions—missed scans, high-risk shrink behaviours—are detected automatically in the flow of shopping, without turning stores into fortresses by locking products away, or using expensive item-level electronic article surveillance (EAS) tags.

For Retail Technology readers, the 2026 takeaway is pragmatic: retailers should stop treating RFID, IoT and computer vision as separate "innovation projects." They are increasingly part of the same operating model upgrade—one aimed at delivering better visibility, faster execution, lower losses, and smoother store journeys in an era of labour pressures and elevated service expectations.