Simbe expands computer vision technology for shelf digitalization in retail

Simbe has announced significant advancements to Simbe Vision, its AI-powered computer vision technology. The company said these enhancements will further solidify its position as the retail industry's most advanced in-store intelligence offering.

With the expanded capabilities of Simbe Vision, the company said these enhancements empower most worldwide retailer banners to move beyond simply identifying problems to implementing AI-driven actions that optimize store operations with unprecedented precision and efficiency.

New computer vision capabilities now available

Simbe said it worked with its retail clients to develop and deploy these new capabilities.

"The most successful retailers are leveraging real-time, AI-powered insights to drive execution excellence," said Brad Bogolea, co-founder & CEO of Simbe. "What sets industry leaders apart is how they're combining cutting-edge computer vision with Simbe's Store Intelligence platform to identify the next best action for their teams. Together, we are enabling faster, smarter decision-making exactly when it matters most, creating a competitive advantage in today's challenging retail environment."

The new capabilities including in Simbe Vision are:

• Low-stock on-shelf detection: Combines volumetric detections and advanced depth sensing to identify products running low before they become out-of-stock, enabling proactive replenishment and preventing lost sales. Simbe said its volumetric-based low-on-shelf detection provides 1.4x more early signals about shelf condition compared to traditional methods.
• Comprehensive shelf intelligence: Detects product spreads, plugs, misplaced items and missing price tags using advanced AI technology, improving planogram accuracy and reducing inventory distortion.
• Expanded inventory monitoring: Adapts to dynamic retail environments by combining mobile and fixed sensors. Simbe’s advanced computer vision fuses data from autonomous robots and fixed cameras, eliminating planogram maintenance and capturing real-world shelf conditions. Experience reduced complexity with higher accuracy and insights that evolve with a changing store, no matter how often layout or merchandising strategy shifts.
• Product recognition & similarity detection: Employs AI-powered visual search for precise product identification across shelves, distinguishing even nearly identical products (such as flavor or size variants) to ensure correct placement and prevent mis-merchandising.
• Shelf-to-stock comparison: Cross-checks on-shelf inventory with backroom stock data to uncover hidden "plugs" that appear fully stocked at a glance, preventing phantom out-of-stocks and improving replenishment accuracy.
• Automated shelf tag verification: Utilizes barcode scanning, optical character recognition (OCR) and proprietary machine learning to detect mismatches between price tags and products, eliminating costly errors.

Scale and performance with Simbe Vision

The company said Simbe Vision delivers continuous, full-store visibility with 98.7% SKU-level identification accuracy and over 99.3% shelf condition recall, which generates levels of precision that significantly outperform typical computer vision systems that average 85-90% accuracy and 75-85% recall.

Simbe said it has analyzed over 60 billion retail shelf images to date, providing real-time data on inventory levels, pricing and product displays. The company expects to process twice as many images in 2025 as in 2024, helping retailers worldwide make quicker and more informed decisions.

"What sets Simbe apart is our ability to deliver full-store intelligence rather than limited audits," said Jari Safi, vice president, artificial intelligence at Simbe. "Our deep learning algorithms provide SKU-level accuracy across our database of more than 18 million SKUs - the largest of its kind in retail =- distinguishing even the smallest product differences to ensure precise inventory tracking and pricing integrity."

Unlike most fixed-camera systems that rely on rigid planogram data, Simbe's Tally Spot uses real-time insights from Tally's daily store traversals. This removes the need for store teams to follow rigid workflows and delivers more adaptive, accurate shelf intelligence.

Simbe's computer vision system scans 5,400 items per hour, providing real-time data that enables store teams to respond quickly to inventory issues. The company said its system detects up to 10X more out-of-stock items compared to manual audits.

Simbe Vision also recommends the next-best actions based on business impact, helping store teams prioritize what matters most and shift from reactive issue handling to proactive execution that improves labor efficiency and drives real results.

Simbe Vision has been deployed by global retailers including BJ’s Wholesale Club, Schnucks, Wakefern, CarrefourSA and SpartanNash. Simbe said its technologies deliver measurable impact through sharper pricing accuracy and faster execution through the accuracy of its advanced computer vision technology.