Challenges in using Visual AI to reduce retail shrinkage

Visual AI could be a true game changer for retail companies as it can detect suspicious activities in real-time and help curb shrinkage. However, retail loss prevention teams are still engaged in discussions on this technology, even as losses are steadily rising year over year. Why are retail companies holding back on largescale adoption of Visual AI?

Will Visual AI involve expensive technology upgrades?

A common concern among loss prevention professionals is whether Visual AI implementation will need them to make expensive upgrades of their existing technical infrastructure.

For instance, retail companies have, over the years, made significant investments to add CCTV surveillance for their checkout counters, aisles, and other vantage points on their store. Would introducing Visual AI mean that these need to be replaced with modern cameras with robust AI capabilities?

Network Bandwidth is another area of concern. With Visual AI needing analysis of graphic and video surveillance feeds, would the implementation involve expensive commitments on network bandwidth? Will this slow down the overall performance of the store’s network?

Server Storage in the cloud is another potential expense that retailers are staring at while evaluating Visual AI solutions. Graphics and videos occupy significant space and could result in making significant enhancements of the storage – within the store or in the cloud.

While considering Visual AI solutions, processing power in the form of CPU and GPU usage becomes a key consideration. The only way to address this need would be to procure and provision expensive server hardware with greater processing power. And when such hardware upgrades are needed across an entire network of stores, the capital expenses can skyrocket.

Visual AI that works seamlessly with existing technologies

At SAI Group, we understand that disrupting your technical infrastructure is not an option. That's why our Visual AI solution is designed to work smoothly with your existing systems.

Here are just a few reasons why:

• CCTV: SAI Group’s Visual AI solution is built to work effectively with existing CCTV cameras that have a minimum resolution of 2 megapixels.
• Bandwidth: By leveraging edge intelligence and analytics, SAI Group’s Visual AI solution uses very limited network bandwidth.
• Storage: SAI Group’s Visual AI solution stores only specific footage, which needs to be actioned upon. This significantly reduces the need for adding in-store or cloud-based storage.
• Alerts on Handheld Terminals (HHT): SAI Group’s Visual AI solution provides apps for the HHTs used in the store, enabling the store personnel to receive alerts and notifications as and when incidents are detected.
• All-in-One Edge Solution: SAI Group’s Visual AI is designed to work with minimal processing power (CPU/GPU) and does not need expensive server hardware.

Advantages of Visual AI for retailers

By integrating SAI Group's Visual AI solution, you can enjoy a range of benefits, including:

1. Reduced Shrinkage: Our solution can help you detect and prevent shrinkage, resulting in significant cost savings.
2. Improved Security: Our Visual AI technology can identify potential security threats, allowing you to take proactive measures to protect your business.
3. Enhanced Operational Efficiency: By automating the detection of anomalies, our solution can help you streamline your operations and reduce the workload of your security team.

About SAI

As a leader in computer vision technology, SAI Group delivers cutting-edge, multi-modal AI solutions into retail environments. Using a unique platform approach, its technology uses existing camera systems to target losses, increase store safety, and underpin operational efficiencies.

All solutions are built from the ground up to ensure the highest levels of security and data protection, respecting the privacy expectations of the public and operating to stringent ethical standards while delivering substantial value to our clients. Globally, SAI monitors millions of transactions per day, protecting the revenues from tens of millions of product sales and hundreds of millions of customer interactions. Its models also accurately identify anti-social behaviour, aggression and violence, helping to de-escalate situations with real-time interfaces to security officers and operations centres.