Rapid Robotics, an industrial robotics manufacturer.

Developed modular robots and AI software that can quickly be integrated into workflows for small- and medium-sized companies, providing access to robotic automation without the hefty price tag.

When a manufacturing or logistics company considers automating with robots, it’s a lengthy—and expensive—process. A systems integrator assesses the company’s need, designs something specific to that use case, and charges by the hour as the solution is developed. The process can take months, costing upwards of $250,000 and presenting other risks as well. “If anything in that use case changes, then typically the machine becomes useless,” Tom Hummel, Rapid’s chief technology officer, tells MIT Horizon.

Rather than develop highly specialized robots with upfront costs, Rapid builds robots with off-the-shelf components, plugs in its software, and leases them to companies using a robots-as-a-service subscription model. Rapid can get its solutions up and running in several weeks, saving time and money. After years of designing and selling robots for manufacturing applications, Rapid is working on new products that specifically focus on third-party logistics providers, which help other companies store and move products that they’re selling. These companies carry out repetitive tasks, and by combining modular robotic hardware with machine learning(ML) software, Rapid’s machines can more quickly be slotted into those workflows.

The robots focus on building and packing boxes and moving pallets, and consist of nimble robotic arms attached to a frame. “We actually use multiple smaller robots to be able to have the fastest possible speed, because ultimately, shipping more goods out of one facility has double or triple ROI advantage,” Hummel says. Robotic pickers using ML can reduce a company’s use of temporary or low-paid human workers, while keeping packing quality consistent. Rapid aims for 99.9% accuracy in its packing and says its lab tests, in collaboration with clients in the meatpacking and third-party logistics industry, are close to achieving that goal.

Rapid Robotics initially aimed to sell its robot-as-a-service model to manufacturing companies carrying out a variety of tasks in order to optimize labor. But the company determined that, at that time, machine learning was “not at the level where you can have repeatable high-volume accuracy” for manufacturing, Hummel says. Several years later, they shifted focus to logistics providers and realized that ML technology had significantly improved. Rather than using one algorithm to train its robots, Rapid relies on many. “With advances in ML recently, there are so many very, very, very high-performing algorithms to use,” Hummel says. “We can generate a solution very quickly.” For logistics companies that are packing up goods for other organizations, robots can accurately pick through products and create a customized box, ultimately reducing labor costs and the time it takes to complete such tasks.

For more on how companies can benefit from ML and other AI techniques, see Benefits of AI. For more about the role of robots in supply chains, see How Robots Are Used Today in Industry.