When searching for an item in store, you might as well flip a coin. Heads you find it, tails you don’t—or at least it has all the wrong details. Maintaining an accurate inventory has long been a challenge for retailers.
Doug Benson, VP Marketing, Americas, at Trax, offers a solution. We recently spoke to Benson about using advanced image-recognition technology to help retailers keep track of their goods and stay competitive as the retail industry transforms.
Q: Despite considerable time and money spent on tracking, counting, and replenishing inventory, retailers still struggle with inventory management. What is your approach to solving this problem?
A: We give retailers unprecedented visibility into the products on their store shelves. By capturing shelf photos using mobile devices, fixed cameras, drones, and/or robots, retailers can improve operational efficiency and increase revenues. Store employees can quickly address shelf issues such as out of stock, planogram compliance, and pricing errors.
Q: What role does artificial intelligence play in the Trax solution?
A: AI is fundamental. Recognizing products, prizes, promotions from photographs of store shelves is a complex process requiring multiple specialized AI engines such as panoramic stitching, fine grain image recognition, context, and other applications.
The key to any AI engine is repetition and volume. Every time the AI engines process new data, they get better at producing accurate and consistent results. Trax has processed more retail images that anyone—more than 7.5 billion images to date. Trax currently processes more than 300 million new images every month, and that number is growing daily.
Q: This is a difficult time for many retailers. How do you help companies stay competitive?
A: One of the primary ways is to increase product availability by quickly addressing out-of-stock items. Globally, retailers report between 4 and 12 percent of items are out of stock. Often, the products are available elsewhere in the store so the inventory can be reallocated. But the retailers need to know when the last products are removed from the shelf. Trax can identify an out-of-shelf occurrence and send alerts within seconds to the appropriate store employees. Just half a percentage reduction in out of stocks can result in significant revenue uplift.
"Humans are simply not effective at analyzing thousands of products and prices. Product packages today are so similar that many look identical to auditors."
"The key to any AI engine is repetition and volume. Every time the AI engines process new data, they get better at producing accurate and consistent results."