Ep 78 | Don't Start With A Whiteboard, Talk To Customers (w/Brennand Pierce)

In this talk

1. 1:45

   From computer science to robotics

   Pierce says his interest in math, hands-on building, and earning money shaped his early choices. After studying computer science, he built software for call centers, then became absorbed in robotics through a Japanese humanoid kit and work at Bristol Robotics Lab. He says seeing a robot move made the work more rewarding to him than software whose results stayed on a screen.

2. 31:58

   Customer conversations beat whiteboard guesses

   Pierce argues that founders should start with a customer problem and then decide what technology could address it. In a Munich entrepreneurship program, he saw PhD graduates brainstorming uses for their research without enough contact with potential buyers. He later took a different approach with a robotics company, spending about a year talking to hundreds of businesses before hiring an engineer or raising money.

3. 38:11

   Lessons from hotel and restaurant robots

   Pierce’s hotel robot company, Robotoise, built a mobile minibar that customers liked, but the economics did not work; a small shop or vending machine in reception was a better answer to guests wanting snacks or drinks. He carried lessons about customer discovery and robot building to Bear Robotics, where Servi’s early deployments revealed that larger restaurants needed larger robots. Shipping roughly 10,000 units helped expose those requirements.

4. 41:00

   Start with tasks robots can reliably do

   Pierce says his current team spent a year investigating applications for foundation-model robotics and return on investment before expanding the team or fundraising. The company is developing a bimanual mobile robot to handle totes and factory parts, including moving components from a stamping line to a powder-coating hook. Pierce says teaching a robot by teleoperation is promising, though a full system for an eight-hour shift was still months from deployment.

5. 51:15

   Deploy in real workplaces and learn

   Pierce cautions that carefully arranged demonstrations do not represent cluttered homes or factory floors. His company aims to place about 100 robots with around 20 customers over the coming year, targeting brownfield sites where automation is limited. Over the next few years, he hopes one model can perform 60 to 70 percent of tasks across settings such as warehouses, including difficult jobs like peeling and placing labels.