How can computer vision and AI enhance loss prevention in retail?

Retail store margins are already severely pressured as a result of online competition. The bottom line of a retailer could also be negatively impacted by shrinkage, or a shortage of inventory compared to the records available in the system. This could result from inventory loss from staff theft, shoplifting, and other factors. It can take place on the sales floor, at the register, or even at the doorway to the store.

Retailers are at risk because more shrinkage results in poorer profitability. According to the National Retail Security Survey 2021, shrinkage hit an all-time high in 2021 and cost merchants 1.6 percent of their revenue.

In addition, in the post-Covid era, while both customers and retailers have benefitted from the push to digitize stores with more contactless and frictionless self-services, like self-checkout, recent advancements like these have opened up the potential for new fraud opportunities.

A Closer Look at Shrinkage

Loss prevention systems’ primary goal is to reduce shrinkage. Cameras, EAS alerts, and RFID-enabled loss prevention are just a few of the improvements in sales-floor surveillance that have been made to keep an eye on both merchandise and workers. However, they hinder retailers’ capacity for a speedy response.

Shoplifting has caused considerable shrinkage, especially within the retail industry. Many thieves formulate different tactics to steal from stores such as concealing items, walking out, switching tags, unwrapping items, inside jobs, etc. These sorts of actions leave businesses’ inventory unmatched, which means physical inventory does not correlate with the store’s book inventory.

In addition, there has been a rise in employee theft, sometimes known as “sweethearting,” in which employees gift products to friends and family without charging them or charging for less expensive items than those being purchased. Many distinct sorts of retail theft can occur at the register, depending on the cooperation of an outside shoplifter and a staff member at the point-of-sale (POS) system.

As highlighted earlier, self-checkout has also increased the risk of theft. Products do not need to be hidden in shoppers’ clothing or in a bag while shopping in-store. Buyers may put the most expensive products straight in a bag at the self-checkout while scanning other items to avoid scanning the pricier ones.

Shrinkage is a serious matter to keep an eye on because it can truly dissolve a business. For example, loss of profit is the main reason for shrinkage due to the fact the products that are being paid to be in the store are not making revenue towards the business, since items are being stolen. The only options that business owners must initiate are either raising store prices or reducing wages for employees. This then brings about a ripple effect because if the prices rise, then the consumers would be dissatisfied and reach out to other competitors with better pricing, and if wages are reduced to employees, the workers would not want to work for a business that does not meet their needs. In the end, a business will eventually fall if shrinkage continues to impact heavily.

Why Aren’t Sensor Analytics Enough?

Sensor analytics is based on the use of physical sensor devices in-store. These work by converting stimuli such as movement or sound into electrical signals which are converted to code and then processed by computers.

New technologies, like AI and computer vision, take sensor analytics a step further, as anything one’s eye can see can be analyzed. For example, sensors may be able to pick up on whether a person has walked into a store and into an aisle. However, computer vision can notice that in addition to whether there’s a dangerous spill in the aisle that could cause harm to a shopper. And it can do so much more when it comes to loss prevention.

Humans are designed to be highly visual and gather information whether it may be daily tasks, environments, or each other all through visual perception. There are quite a few limitations that humans have that AI and computer vision do not. For instance, humans tend to get distracted or become fatigued; however, AI and computer vision can be installed anywhere and provide 24-hour vision daily.

Sensor analytics has its benefits of monitoring and detecting anomalies, but the issue is with the data collected. Since sensors are always on, there are challenges when it comes to collecting, storing, and interpreting all the amounts of data created. When it comes to computer vision and AI, depending on the input of algorithms, it can capture any video feed in real-time and based on the video feed, it can collect, interpret, and create analysis instantaneously. AI technology formulates dashboards, charts, etc. to make sure the business runs smoothly and efficiently. Even though sensor analytics can be helpful, it is not enough, that is why AI and computer vision is beneficial to implement within in-store businesses.

Use Computer Vision and AI to Reduce Product Loss and Checkout Fraud

As a result of the aforementioned expanding problems, retailers are investing in new technology to combat retail shrink with improved loss prevention measures at the front of the store. Innovative technologies integrating AI and computer vision are now readily available to assist businesses in reducing theft, combating shrink, and eventually improving inventory management. They also offer the added benefit of improving the customer experience by preventing unplanned stock-outs caused by unforeseen thefts.

To combat checkout fraud, both at manned registers and self-checkouts, integration of data from item-level tracking with computer vision and POS is necessary. By comparing item-level counts to POS-generated counts, associates can detect fraud and take fast action. With the help of AI and computer vision, retailers can make checkout processes smarter, which reduces theft and improves inventory control.

What could this look like in practice? In order to identify and prevent theft by shoppers, a retailer could add a camera to the existing checkout lanes and use AI to compare the number of products scanned. The camera deployed sees the items being scanned as a customer scans them at the POS system. The integrated POS system receives the entire item count right away when it is generated. The POS system and the video camera are connected, and after the items are scanned, the POS system adds the camera-generated count to the overall count.

If the two numbers don’t match, a system can be set up to warn the staff member in charge through a dashboard, POS alert, or mobile alert of any potential theft or improper billing. It enables store staff to intervene and provide assistance prior to a transaction being executed, allowing the manager to process the transaction manually.

Thanks to computer vision and AI, retailers now have the opportunity to make things right without having to accuse customers of stealing. Customers can be told about the system problem and given the choice to pay for the unscanned items by store staff. In the end, this will aid in behavioral change, which will reduce shrinkage and save money.

What Should Retailers Look for in a Solution?

To minimize costs and maximize current technology investments while speeding time-to-benefit, retailers should consider offerings that are hardware-, cloud- and system-agnostic so they can use their current video and data infrastructure. They should also seek solutions that are scalable and can meet business needs however large they grow.

Retailers should also research solutions that offer dashboards so retail staff can gain visibility into all needed metrics in one easy-to-access location. Platforms that provide real-time alerts, as discussed above, enable retailers to identify issues (and opportunities) in real time so they can immediately act as needed for their business. In addition, with solutions that deliver data-driven predictive insights, store managers can implement changes that not only protect the business in the moment but also in the longer term.

In Summary

According to the National Retail Security Survey 2021, retailers are making significant investments in technology to fend off the rise in organized retail crime. Investing in a loss prevention solution that uses AI and computer vision would quickly and positively reduce product loss and related revenue leakage. The technology can also enhance the consumer experience, which is another advantage. AI and computer vision are advantageous to both retailers and consumers alike.

 

About the author: Rohan Sanil is the CEO and co-founder of Deep North, an intelligent video analytics company. He has over two decades of product, business, and entrepreneurial leadership in the video analytics space. He previously founded Akiira Media Systems, Atstream Networks and Tri-Cad, where he was instrumental in product development, business development, and raising capital. Rohan holds an M.S. Degree in Management Science from the University of Dayton, Ohio, and a B.S in Mechanical Engineering from Karnataka University, India.