Melbourne start-up Black.ai has been trialling its system of connected sensors and cameras at a major supermarket chain (co-founder Keaton Okkonen won't say which one) with the aim of eventually allowing it to dispense with cashiers and in-store detectives, and allowing it to automate functions like ordering new stock from a distribution centre.
While the technology to run a cashier-less supermarket already exists -Amazon is reportedly planning 3,000 of its no-checkout AmazonGO stores by 2021- Mr Okkonen claimed Black.ai's ability to share data between its "seeing computers" would bring down the cost of store automation and further accelerate its deployment.
"If you've got a fleet of robots that can all plug into an information layer and share what they're seeing, you're able to place cheaper, lower-resolution sensors in each one," he said.
Black.ai is also pointing its algorithms at a growing problem faced by supermarkets since they replaced many of their human cashiers with self-service checkouts – "stock bleed", as Mr Okkonen diplomatically called it, which occurs most among unpackaged, non-barcoded items like fruit and vegetables.
"In the trials we've done, we've already successfully identified somebody grabbing an organic avocado, then entering it as a regular avocado at the checkout," he said.