Nov 24, 2021 - Technology

Meet the robots that will help pick your Black Friday orders

Animated GIF - Find & Share on GIPHY
Nimble Robotics machines picking goods at an e-commerce warehouse. Credit: Nimble Robotics

As more of our goods and gifts are being purchased online, robots are set to play a bigger role in e-commerce warehouses.

Why it matters: With retailers struggling to meet the demand for workers in a tight labor market, smarter robots that can adapt to the often chaotic conditions of a warehouse could help close the gap.

When you place an online order this holiday season, it will need to be fulfilled by a retailer, which means that good has to be physically picked and packed at a warehouse.

  • It's usually been humans doing the work — as of August, nearly 150,000 more people were employed in e-commerce warehouses than before the pandemic.
  • While the often monotonous work might seem perfect for robots, the unpredictable environment of a warehouse and the varied nature of different goods — think everything from a tube of toothpaste to a crystal figurine — has "been considered un-automatable," says Simon Kalouche, the founder and CEO of Nimble Robotics.

Yes, but: Retailers — especially brick-and-mortar companies that are moving increasing amounts of their business to e-commerce — are having trouble finding enough workers, opening the door to automation.

  • "More companies are turning to robotic solutions not to get ahead of the e-commerce boom but just to keep up with their existing orders and have hope for the holiday season," Kishore Boyalakuntla, VP of products at robotic supply chain provider Berkshire Grey, told ZDNet this week.

How it works: Nimble Robotics is using computer vision-enabled AI — and human helpers — to train robots to pick and pack a wide variety of e-commerce goods.

  • "We use an algorithm called imitation learning, where you use a human to remotely operate a robot to accomplish a certain task," says Nimble's Kalouche. "If you do this hundreds or thousands of times over various tasks, you can train a neural network to say, 'I've seen this and I know how to pick it up.'"
  • As Nimble's robots are deployed in a warehouse, human remote operators are still available to help them through any unexpected obstacles, but over time, "you go perhaps from 90% autonomous to nearly fully autonomous."
  • Nimble's hybrid approach — as well as its subscription model, which charges customers per good picked — makes it easier to integrate automation into a warehouse without substantial upfront costs.

The bottom line: Robots won't be fully replacing human warehouse workers anytime soon, but an automated e-commerce Santa Claus may not be too far off.

Go deeper