Simbe Robotics announced this week integration of RFID and machine learning into its Tally shelf-scanning mobile robots. Tally now captures more than 700 RFID product tags per second with higher than 99% accuracy, according to the company.

So Tally can now determine inventory levels for products that use RFID or that don’t sit on a standard shelf. Brad Bogolea, Simbe Robotics CEO and co-founder, tells The Robot Report, having RFID capabilities opens the door to new markets Tally couldn’t serve when it launched in 2015.

“We were originally focused on packaged goods, so adding RFID to our product was a natural evolution for us,” he says. “Higher-end retailers that sell consumer electronics, clothing and home decor were interested in us adding RFID back when we launched.”

This is Tally’s first major spec upgrade since Simbe announced a partnership with Softbank Robotics America in mid-April 2018. Softbank is helping Simbe commercialize Tally in Europe, Japan and North America. If all goes well, it will then help Simbe commercialize Tally more broadly.

Must-Read: Can robots save retailers from an apocalypse?

This is good timing for Simbe as there’s increasing demand for RFID-capable mobile robots. And the competition from Bossa Nova Robotics, Badger Technologies and others continue to stiffen. Bossa Nova just acquired computer vision and facial recognition software company HawXeye. We’ll have more on that soon.

Tally’s machine learning capabilities improve its navigation through stores and around obstacles. Bogolea says the machine learning also improves Tally’s accuracy when scanning RFID tags or for products on shelves.

“Brick and mortar retail is changing rapidly. Innovative technologies, like RFID and computer vision, are starting to become more mainstream as retailers think about how they can optimize in-store operations,” says Durgesh Tiwari, Simbe’s Vice President of Research and Development, Hardware.

Simbe hopes Tally can help solve the $1.1 trillion in annual retail losses retailers face due to out of stocks, over stock of merchandise, and product location errors.

Since its unveiling in November 2015, Tally has been deployed by 11 international retailers and has safely navigated 4,500-plus kilometers in-store alongside customers and employees; captured more than 32 million shelf photos, analyzing over 150 million products and shelf tags, and completed more than 7,500 hours of fully autonomous operation.