GreyOrange Acquires Site Bionics: Computer Vision Moves RFID Toward Broader Store Autonomy Offering

In October 2026, GreyOrange announced the acquisition of Site Bionics, an Artificial Intelligence (AI)-powered platform that uses retailers’ existing security cameras to model the store and the people moving through it. The Site Bionics  technology will initially be added to the gStore retail platform, but is later expected to extend into warehouse operations through GreyMatter. Site Bionics is a small startup—one of its co-founders, Adam Blair, was the founding Chief Technology Officer (CTO) of RADAR, a company known for its always-on overhead Radio Frequency Identification (RFID) solutions, with a product that, until recently, also included Computer Vision (CV).

The combination of GreyOrange’s always-on overhead-first strategy with Site Bionics’ CV technology is, therefore, no coincidence. The synergies between these technologies have long been explored, perhaps earliest by Zebra Technologies’ SmartLens solution in 2017, and continue to be assessed. gStore already uses item-level RFID data to understand what products are in the store and where they are, while Site Bionics adds the visual context around those products: how shoppers and staff move, where they linger, what they interact with, and whether displays have been executed as planned. GreyOrange is positioning the combination to create a more complete view of the physical store without deploying new camera hardware. It has highlighted applications including shopper behavior analytics, visual merchandising, and faster loss investigations by linking evidence from different sources.

Retailer interest in Closed-Circuit Television (CCTV) Artificial Intelligence (AI) solutions is increasing, both standalone and in combination with RFID. Greater adoption of overhead RFID will unlock additional value from the relationship between the two technologies, initially as video streams provide real-time visibility where RFID only provides snapshot views; but increasingly as more frequent scans from always-on overhead RFID solutions enable real-time product and visual information to provide more valuable insights together. Retailers are increasingly thinking about how to use AI solutions on video streams, initially for separate workflows to RFID, but with an eye on assessing how to merge these data sources. AI-based CCTV solutions are still for early adopters, with broader based adoption and combination with RFID likely not to be seen for a few years. However, suppliers must position for this market and start learning and demonstrating the benefits to customers.

Existing CCTV provides an easier route to deployment than deploying another dedicated sensor estate. Most retailers have store cameras, even if they were installed for security rather than analytics. Site Bionics, like other retail video analytics suppliers, gives GreyOrange a route to extract more value from this installed infrastructure and introduce computer vision without requiring retailers to fund another hardware rollout. This broadens the opportunity beyond overhead RFID-led retailers. GreyOrange can sell standalone vision applications, or add vision to handheld- or overhead-led RFID environments with different tiered offerings. This allows retailers to adopt in a modular way in accordance with operational readiness, Information Technology (IT) readiness, and Return on Investment (ROI).

The acquisition is more than a bolt-on CV capability. It reflects a broader change in how GreyOrange wants to position gStore: not as an RFID platform that reports data, but as a store intelligence and execution layer that helps determine the next action—moving up the RFID stack to create higher-value solutions. The “next best action” terminology provides recommendations based on the data, instead of focusing on dashboard and data reporting, and can market how platforms are evolving. Few suppliers have this capability today, historically focusing on simpler reports and dashboards, but more are positioning themselves along this vein, with growing roadmap focuses on AI and analytics. The differentiator will be providing a unified environment to ingest and analyze data from multiple sources, and turning data into actions for store associates, managers, asset protection teams, or planning teams.

The acquisition could change the traditional commercial model for RFID platforms. CV and sensor fusion create opportunities for premium applications packaged around a tangible outcome. This moves the commercial conversation away from pricing primarily around readers, stores, or basic RFID workflows and toward specific operational outcomes; these repeatable patterns can facilitate adoption by retailers and provide clear benchmarks to assess a pilot. The central question is whether retailers will pay separately—and how much—for these applications, or expect them within the core gStore platform, in an RFID market known for increasingly low-value software licenses.

The acquisition is also defensive against competitors such as Nedap, ClarityRFID, Checkpoint, and other RFID software vendors. GreyOrange differentiated through its first-mover advantage in overhead RFID, but as other software vendors develop their offerings, it needs additional value to avoid joining the traditional RFID platforms, which cannot command a similar license price. Adding CV gives GreyOrange another differentiator, but, as with always-on overhead RFID capabilities, the advantage will only last if the company turns additional sensor data into repeatable products and measurable outcomes faster than competitors can build, buy, or partner for similar capabilities.

Adding capabilities from different sensor sources reflects a broader shift toward “store autonomy” solutions, instead of disparate point solutions such as RFID, CCTV, people counting, Electronic Shelf Labels (ESLs), kiosks, and digital signage operating in silos. These larger business cases expand the value available beyond the traditional Automatic Identification and Data Capture (AIDC) integrator approach of deriving most of its value from hardware resale, because hardware is considered a quick-sell, compared to broader automation projects. Managed Service Providers (MSPs) will play a more important role, but must change their attitude and approach, moving beyond the quick wins of hardware resale and managing more complex integrations and customizations.

The growth of video analytics in retail revives the question of on-premises versus cloud processing. Processing video data in the cloud is extremely costly; processing on individual cameras is a first filtering stage but limits the context from the broader autonomous store. The concept of the edge box becomes relevant: not necessarily a full local server, but a layer that makes integration and local processing easier. Retailers have sought to move systems to the cloud, and this will remain important for data lakes like Snowflake, which allow centralized processing and feed enterprise systems; but cloud platforms will increasingly be complemented by local edge processing. Always-on overhead RFID solutions frequently process data at the edge to limit data sent to the cloud, an approach growing more important as scanning frequency increases; video growth will accelerate the same trend. “Edge versus cloud” is not binary: GreyOrange and other ISVs must determine inference at each layer, based on latency, bandwidth, privacy, and cost. Camera-level processing may filter footage and detect events; a local edge layer may correlate video with RFID and other store systems; and the cloud may support cross-store analysis, planning, and forecasting.

Video brings privacy and governance questions to the forefront. Today’s RFID platforms frequently use proxies, such as people counting sensors or basket groupings inferred from always-on overhead RFID reads, to generate shopper insights anonymously, mitigating privacy concerns. As video use grows, clarity around data capture, retention, and inference will be crucial, and will vary by use case and region. These questions will affect whether deployments scale, and across which use cases.

ISVs must consider not only how to include more types of sensors within their platforms to meet broader store autonomy solution requirements, but also how to productize and scale these solutions. Two questions emerge: 1) how can the platform translate data into actions, so that large volumes of information become meaningful rather than creating data overloads; and 2) how should ISVs partner with integrators and MSPs instead of owning customer relationships directly?

MSPs are under-recognized actors in retail technology, but ISVs will increasingly realize that the value of their solutions—and how much they can charge—will depend on MSPs’ ability to use their platform to integrate broader automation solutions. Conversely, ISVs are the glue for these MSPs and integrators. As their platforms reposition around store autonomy solutions, their approach of owning customer relations will shift and increasingly rely on customization by MSPs to help retailers gain greater value from deployments. MSPs will increasingly rely on ISV platforms to ingest disparate data sources as a store autonomy platform rather than an RFID-only platform. The ISV will provide the common data and application layer, while the MSP will sell customization and ongoing solution management. The Site Bionics acquisition demonstrates the growing relevance of this ecosystem question. The broader the platform becomes, the more important partners will be in turning ISV capabilities into retailer-specific outcomes.