San Francisco, CA, USA, July 5, 2022. Shoppers around the world have experienced going to the store to buy their favorite shampoo or candy bar only to find the shelf empty. Because of inflation and high rates of employee turnover, retailers are struggling to keep their shelves stocked. Supply-chain issues caused in part by the COVID-19 pandemic have added more challenges in keeping shelves stocked and increased the product-unavailability rate from 5 percent to 15 percent during the past three years, according to the Consumer Brands Association.
Startup Wisy developed an AI platform to make it easier for stores to track whether there are available products that haven’t yet been put on display. It uses image recognition to detect which items need to be restocked.
“We are not only solving a customer-experience problem but also a sustainability problem,” says IEEE Senior Member Min Chen, Wisy cofounder and CEO. “All those products that are not sold because they were not displayed get thrown away. WisyAI enables store employees to quickly get information about the stock, reduce losses, and sell [products] more effectively.”
Chen was featured on the cover of Forbes Centroamérica‘s July/August issue.
The startup, headquartered in San Francisco, received the 2022 Startup of the Year Award at the Cloud Wars Expo. Wisy also won this year’s CCU (Compania Cervecerias Unidas) Chile’s Innpacta Global Open Innovation Challenge. The competition is for startups that have designed technologies for the consumer goods and retail industry. Wisy is piloting its AI platform with CCU.
Unreliable Internet Connection
Chen, a software engineer, founded her first company, Alcenit, in 2006. The consulting business worked with companies in the banking, retail, and oil-and-gas industries to help them manage their IT department. One of Alcenit’s clients—a supermarket chain—was having problems collecting reliable ground-truth data, or data collected at scale from real-world scenarios.
Chen discovered that the retailer wasn’t the only client struggling with product data collection.
“We had customers in the oil-and-gas industry, as well as construction companies, dealing with similar problems,” she says. She and three of her coworkers decided in 2017 to launch Wisy to develop a solution.
Wisy decided to develop the technology for supermarkets and other retailers, she says, because the tracking process they used caused problems that were “big enough, repetitive enough, and complex enough to solve using AI.”
To learn what the issues were, Wisy conducted on-site observations and interviewed employees about their process for collecting product data.
“Usually consulting companies do this,” Chen says, “but we felt it was important for us to do it as the platform developers.”
They found that retail employees manually tracked and verified what products were available in the store, using cloud-based platforms. Employees counted each product and recorded information about the item, such as pricing, on a platform.
Manually counting the products and filling out the information was a slow process, and employees were prone to making errors, Chen’s team found. The cloud providers that the retail stores used often were unreliable, she says, and the Internet connection was regularly slow or unavailable, especially in stores’ basements and warehouses.
“If employees didn’t have a good Internet connection, they wouldn’t be able to see the results until the connection was reestablished,” she says. “If their connection was slow, then getting the results was slow.”
Adding to the stores’ challenges were worker shortages.
“It is quite a challenge for our clients because they have a high number of vacancies, so they have to do more work with less people now,” Chen says. “Training employees on how to do in-store placement and tracking of products also takes time—even more so with high numbers of employee turnover.”
Keeping Shelves Stocked
Chen says she knew Wisy’s platform had to be “available, reliable, fast, and flexible.” It also had to work for large companies as well as mom-and-pop shops, and without an Internet connection.
Wisy developed a platform that uses AI to process data regardless of whether there is access to the Internet. The service itself, which works on a mobile phone or tablet, has the power to process data and produce results in milliseconds no matter where the user is in the store.
“With Wisy’s platform, companies can be assured that their teams have a reliable tool to support their work,” Chen says.
Images of each shelf and product, as well as their bar code, are uploaded to the platform. Store associates take a picture of a product on display, and the AI records information based on the photo. It takes less than one second.
Wisy AI can work as a stand-alone system or be integrated with the store’s previous product tracking platform.
If there is a shortage of a certain product in the shop, Wisy AI notifies a store associate. Should the platform detect that the product is out of stock or the inventory is running low, it prompts an employee to order more.
It also can track when products were delivered—which is helpful if a customer is searching for a specific shampoo, for example, but it isn’t on the shelf. Employees can use Wisy AI to find out whether the shampoo is in the store’s storage area or is, in fact, out of stock.
The platform uses the data it collects to predict when items will be delivered and when they will go out of stock, based on their popularity.
Wisy released the platform with offline capabilities in April.