Challenges and Surprises from the Robotics Startup Trenches
September 19, 2019      

Building a robotics company is certainly not an easy task – otherwise, everyone would be doing it these days. While the forecasts and predictions all point to growth and dollar signs and huge markets, getting to that point requires lots of patience, knowledge, and solving lots of challenges – as witnessed by several robotics companies in recent years that have not succeeded.

A panel session at RoboBusiness 2019 will address some of these topics in “Building a Company from the Ground Up: Lessons from the Pros.” Moderated by Schematic Ventures Partner Julian Counihan, the panel includes Erik Nieves, the founder of Plus One Robotics; Josh Lessing, the co-founder and CEO of Root AI; and Cyril Ebersweiler, general partner at SOSV.

Robotics Business Review spoke with Nieves and Lessing about some of the issues that rose during the early days of each of their companies.

Plus One Robotics develops advanced vision capabilities with high-performance manipulators as part of its PickOne system, an automated, mixed parcel singulation system. The company also offers Yonder, a remote robot management solution, and CheckOne, an automated QA system for parcel induction.

Root AI develops a line of robotic harvesting and farm optimization technologies to help pick crops more effectively. This includes real-time detection of whether fruits are ripe and ready to be picked via computer vision, grippers that can reach into the plant to pluck a single fruit without leaving a mark, and intelligent motion that has the robot moving in the correct path.

Prepare for long journey

For people interested in creating and growing robotics companies, Nieves said a key message is to be prepared for a long journey. “If you’re trying to do the Silicon Valley Shuffle – Seed Round, A Round, B, and then exit – pick a different industry,” said Nieves. “Robotics requires a level of commitment from the entrepreneur and founder if you’re going to do this the right way.”

Erik Nieves robobusiness preshow article

Erik Nieves, Plus One Robotics

The long journey involves things like the complexity of the systems being deployed or sold to an industry, as well as the number of decision-makers that have to sign off on a deployment, Nieves said. “If you’re a startup or an investor that’s thematic and wants to be looking at supply chain and warehouse automation, they’re going to have to understand that there are realities of lead time,” said Nieves. “Robots don’t replicate the way software does – you don’t just throw another five licenses on the credit card and all of a sudden there’s more instances. It doesn’t work that way. The duration between the pilot phase of the adoption and the subsequent wider rollout is longer.” “Even so, I’m bullish on the opportunity this industry presents and am glad to have 30 people at Plus One pulling with me to make it happen.”

In addition, the number of stakeholders involved in a decision can get higher and higher, he added. “Think about the people involved – operations, infosec, legal, corporate R&D,” said Nieves. “There’s a lot of folks you have to convince that this is the right approach. That takes time, but then you’re good. Especially the bigger the firm, the longer it goes.”

Surprises in the startup journey

Both Nieves and Lessing said they experienced surprises during the course of building their company, but that it wasn’t around difficulties with fundraising or any technologies.

Josh Lessing Root AI startup preshow article

Josh Lessing, Root AI

For Lessing, it was learning how to correctly grow a tomato plant. “In order to test our robots, we had to replicate our customers’ farm, which has given us a newfound appreciation for how much knowledge and skill it takes to be a grower,” said Lessing. “We have since hired a plant scientist.”

Nieves said one of the surprises he found in the early days of Plus One was the sluggishness of the industry to “adopt technology writ large.”

“Everywhere we put a robot almost to a rule, was the first arm they’ve ever had,” said Nieves. “You know that going in intellectually – that’s why we wanted in this market: it was nascent, it was emergent, it was going to be big. It meant we were going to be putting in onesies and twosies at a lot of places to get them started and seed the market. But coming from automotive, where you’re deploying robots by the dozen – it’s a bit of a shock to the system.”

Beyond the end users, Nieves said he still sees a lot of roboticists and robotics companies that get “way too enamored with the technology” they create. “I still see a lot of people that want to do robot trajectory planning,” said Nieves. “Robots have been very good at going from A to B for 40 years. All you need to do is tell it where B is, and it can take care of the rest. But roboticists love trajectory planning, and too many of them get caught up on this.”

Nieves said this was one of the reasons that everybody went after the each-picking problem initially in the supply chain. “It’s the sexiest problem in the supply chain to a roboticist,” said Nieves. “Look at all that variability, look at all these different grasp plans we’re going to have to do, look at all of the different perception problems and different materials present. That’s what they do instead of stepping back and looking at the value chain of the labor in the building and then saying, ‘What is the most effective deployment of my capital and expertise in removing some of their costs. They don’t think that way – they think of the tech first. Too many of them are still doing that.”

Takeaway messages

While Nieves said he hopes attendees walk away with an understanding of how long the startup journey in robotics will take, Lessing had a different takeaway message.

“The most important part of building a robotics company is elite product management,” said Lessing. “Always, always, always close the loop with your customers. Ultimately, it’s their opinions that matter.”