Bossa Nova Robotics today announced a $29 million round of funding to scale up production of its mobile inventory robots for retailers and grocery stores.
New investors in the Pittsburgh-based company include LG Electronics and China Walden Ventures. The funding round included existing investors Cota Capital, Intel Capital, Lucas Venture Group, and WRV Capital. With the latest funding round, Bossa Nova has raised more than $70 million, the company said.
It also announced Flex as its official manufacturing partner to accelerate production of its robots to meet the needs of increasing customer demand. Martin Hitch, chief business officer for Bossa Nova, said the goal is to scale up manufacturing to 200 to 300 robots per month.
Bossa Nova’s mobile inventory robots use computer vision and deep learning algorithms to scan the shelves of a retailer, provide data to retailers on current inventory levels, and notify associates when a product is out of stock. The company has partnered with Walmart for four years, with robots currently more than 50 Walmart stores, Hitch said.
At the recent Walmart shareholder’s meeting, Bossa Nova’s robots were mentioned as a part of the technology that is transforming the in-store experience at Walmart. Information from Bossa Nova’s robots flows to the Downstock app, which shows associates which items are out of stock and where to get them in the store, such as an overstock area or loading dock.
John Crecelius, vice president of central operations for Walmart U.S., told media at the shareholder’s meeting that different technologies were working together to enable faster, smarter work for employees, but also faster shopping for customers.
“All of this comes together to improve our accuracy and availability for our customers, and to improve the jobs of our associates,” Crecilius said at the meeting.
Robot data provides process adaptation
Hitch said the original goal of the mobile inventory robots is to “solve for out of stocks – when a product sells out, make sure it’s restocked as fast as possible.”
However, data from the robots is also now changing other processes in the Walmart stores, such as prioritizing how workers unload trucks, as well as how they speed product from the truck to the selling floor, Hitch said.
In addition, the data is used to optimize pick-and-pack processes for online orders that are fulfilled from individual stores. The mobile inventory robots can help locate products with sub-centimeter accuracy for items that may have been shelved in a different location or moved to a different part of the store.
Martin Hitch, Bossa Nova Robotics.
“So what started up as one use case has morphed into many different use cases across the retail enterprise,” Hitch said.
Computer vision and deep learning algorithms are at the heart of the data capture technology for Bossa Nova. The typical store the size of a Walmart has between 150,000 and 180,000 SKUs, Hitch said.
“The ability to recognize those products and understand the state of those products on the shelf is not a trivial exercise,” he said.
Larger objects, such as big boxes, can be identified quickly and accurately, but when it comes to smaller items, such as spices that are packaged in very small cylinders and moved around on the shelf, identifying them accurately becomes more challenging, he said.
However, the library of images being scanned by the robots is continually growing, enabling the algorithm to learn which products are which more accurately – even being able to distinguish between a can of Campbell’s chicken soup and Campbell’s chicken noodle soup, which look almost identical.
The mobile inventory robots also collect data on things that prevent scans from happening, such as a cart or customer blocking the shelves. That data can be used to improve the robot’s actions, such as whether to stop, navigate around the obstacle, or notify the customer to move.
Workers welcome mobile inventory robots
Data collected by the robot during inventory scanning has also changed other processes for Walmart. Credit: Bossa Nova Robotics
For workers, the mobile inventory robots have helped eliminate one of the more mundane tasks of their jobs – the scanning of inventory to determine out-of-stock items.
“No member of staff likes to scan shelves,” Hitch said. “It’s a mundane and repetitive task, walking down [an aisle] with a handheld scanning gun and pointing it at gaps on the shelf. We haven’t met a single person in the world or organization who enjoys that.”
Second, the data from the robots help make it easier for the associates to restock missing items, making their day-to-day work more productive, Hitch said. “The associates have been incredibly supportive,” he said.
Hitch couldn’t disclose whether Walmart plans to deploy more mobile inventory robots in additional locations, but he acknowledged that the companies’ agreement was not exclusive. Bossa Nova is working with other retailers, Hitch said, but he couldn’t announce any partnerships or agreements at the moment.
Bossa Nova chose North Carolina-based Flex partially for its ability to build high complexity products at lower volumes. Having manufacturing in the U.S. makes it easier to troubleshoot potential problems, Hitch said. In addition, the company’s source materials network was a key factor in the decision.
“Flex is the No. 2 contract manufacturer in the world,” Hitch said. “The power of their network to secure materials at scale and competitive prices, and manage that supply chain on our behalf, is obviously a significant factor.”
