SAI, the leading active intelligence solution for stores, has announced that it has been awarded a U.S. patent for its Visual Language Model (VLM) technology. 

Its VLM underpins the store intelligence running through its SAI One platform, which allows retailers to turn existing cameras into continuous store intelligence, moving beyond alerting to the next best action.

From passive images to operational insight

Issued under the U.S. Patent and Trademark Office (U.S. patent number: 12694682), SAI’s VLM turns traditionally ‘passive’ store surveillance systems into an active, operational platform.  Taking visual data flows generated from in-store cameras – from in-aisle to shelf-edge and point of sale to point of entry – it leverages Vision AI blended with GenAI to transform image-based data into real-time store intelligence.

Building a comprehensive, contextual picture of how stores actually operate, it then surfaces meaningful operational signals, insight cues and timely staff alerts that drive estate-wide performance, from Loss Prevention to Operations, including shopper flow, queue management and health & safety, to CX and Retail Media, such as heat mapping and dwell time.

The patent recognises the solution’s unique ability to create reusable store intelligence across multiple functions, leveraging its multimodal AI video-language architecture to convert time sequences of camera frames into contextual, machine-readable insight. 

An intelligence layer for stores

By acting as an intelligence layer that translates isolated actions into context, the solution turns single visual images into structured operational data flows, which integrate with in-built AI to initiate the appropriate – and most profitable – next action for retail teams. 

This context-first approach moves beyond generic rule-based triggers, learning spatial and temporal relationships across video to understand activity in the context of each store’s unique operational environment and trading landscape.

Rather than being confined to a dashboard or single use case, the platform actively and agnostically connects store data feeds from any device or system – from POS and headsets to handheld devices and CCTV – to turn visual cues into prioritised, profitable actions.

“Running stores has never been more complex, meaning retailers need unified, end-to-end store intelligence solutions that connect the dots – between systems, processes, people and functions – to optimise their increasingly connected in-store environments.”

Abhijit Sanyal, Chief Technology Officer, SAI

“It also means retailers need solutions that break away from generic rule-based actions and instead use AI to hard-bake contextual intelligence into store operations,” Abhijit continued.   

Beyond Computer Vision

“We’ve engineered our Visual Language Model to move beyond Computer Vision alone, combining Vision AI with GenAI to understand not just what is happening across the store, but why it matters. Securing this patent validates that innovation and reinforces our commitment to pushing the boundaries of intelligent retail operations,” he added.

SAI Group will be showcasing the latest advancements to its VLM and its suite of store intelligence solutions at NRF Europe (15 – 17 September, Paris Expo Porte de Versailles). Exhibiting on Stand #B-044 in Hall 6, SAI will highlight how retailers can harness innovation to run and protect their stores, using visual data and AI to drive estate-wide performance.

To connect with SAI at NRF Europe, or to find out more about SAI’s VLM, the SAI One platform or its suite of solutions, visit: www.saigroups.com.