April 10, 2020
That tech entrepreneur Tom Stevens’ work life revolves around a soft immortal puppy is an outgrowth of heartbreak. His big idea came in 2011, soon after his mother was diagnosed with Alzheimer’s and had to give up her dog for safety reasons. Dealing with her sadness was harrowing and so was recognizing her loss of depressive-reducing benefits that research shows comes with spending time with pets.
Stevens started looking for substitute companions. The market was bleak, littered with goofy stuffed cats and one ultra-expensive robotic seal that soothed dementia patients with batting eyelashes and electric warmth. Stevens, who had more than 30 years working in tech and had recently sold his e-discovery software firm, wondered if he could do better.
With his mother as his first official product tester, Stevens spent the next few years reaching out to hundreds of the country’s 5.6 million senior citizens suffering from Alzheimer’s, honing their preferences into a robotic lap-Labrador who responds to voice and touch. Five prototypes and one Master’s of Science in Management from Stanford GSB later, Stevens and his six-employee startup, Tombot, are on a quest to ease the loneliness, anxiety, and frustrations that often plague dementia patients.
Stevens, whose company launched in 2017 and is based in Valencia, Calif., discusses the challenges of breaking into the medical device industry, making a mechanical creature lovable, and unlocking a potentially vast new market.
Your mother inspired the genesis of your company, but what about its name?
My mother couldn’t remember what she had for breakfast on a given day, but she always remembered that I was working on robots for her — she tested all of our early prototypes. It was what she wanted to talk about and it helped change the topic from the emotionally charged memory of “where’s my dog, why can’t I have my dog” to “do you have another robot for me.” It became a force of positive energy in our relationship, and she said one day, “You should name it Tom.” I said, “Maybe not Tom exactly,” and it became Tombot. She loved that.
Your past career was working in software and legal IT. How did you get your robotic/medical device upstart rolling past the idea phase?
It wasn’t until I was ready to enter Stanford GSB that I was convinced there was both a big opportunity to work on and that I was capable of working on it. I went to GSB with the intent to launch this company. I had done a tremendous amount of research and was well equipped intellectually to do this, but I was changing industries and just didn’t feel confident that I was capable of leading a company that wasn’t exactly what I had been doing previously.
How did you approach R&D into this new-to-you sector?
I’m a high-tech guy. I like computers and I’ve been a fan of robots for a long time. So I entered this product design process as free as I could be. It’s human-centered design where we’re focusing on the customer. Initially that was my mother; ultimately it became hundreds of seniors with dementia. We came to realize that seniors with dementia prefer things that are animated versus things that are static — something robotic over a doll or stuffed animal. They also have a very strong preference for things they’re already familiar with, like a dog versus a mythical creature or space alien. Lastly, they have a strong preference for things that are realistic.
So no dragons or aliens. But what about baby harp seals, which a longstanding Japanese competitor built its business around in 2005?
I learned about PARO, which sells for $6,000, in 2011 when I was looking at products for my mom. Two things bothered me — the price and why a seal? Why would this be something that’s appealing?
Well, a cute robotic seal avoids the threat of uncanny valley, a theory that describes our strong aversion toward things that appear almost realistic but aren’t quite right.
Before I started this venture, I had real skepticism that the theory would apply to robotic animals, though it’s likely at the heart of why PARO went in that direction. The theory is from a Japanese roboticist who speculated that as humanoid robots became more realistic, we would prefer them until they got fairly close to looking real, and then they would creep us out. While uncanny valley is a real phenomenon — it’s also been part of Hollywood since the Disney days of early animation — my hypothesis is that it wouldn’t be true cross-species, and we tested that.
You also hired a well-known animatronics firm to make sure your robot, which ships with the name Jennie, appears as charming as the breed she represents.
My co-founders and I — we’re not artists. And we knew from research out of MIT that if robots are appealing at all, they have to be cute and adorable. We weren’t trying to do horror versions that might trigger uncanny reactions. We considered more than 40 animatronics firms, and Jim Henson’s Creature Shop was by far the best from a skill set standpoint.
What makes it the right time for Jennie?
It’s less about the eldercare market and more about the timing of the evolution of technology. PARO is a great example of a company that was too early. Not because the senior care market wasn’t ready, but the technology wasn’t there to support it. They did their R&D in the 1990s when there was little they could license — they had to invent it. They did a great job but it was extremely expensive, more than $15 million in R&D costs. Thanks to the advent of things like the smartphone and 3D printing, we’ll make it to market for one-third of that.
Yet figuring out how to manufacture Jennie to scale cost-effectively is still a puzzle you’re working out.
There are a number of questions that have to be answered through that process, though none are showstoppers. For example, the single most expensive component in the product is the servomotor that drives her motion. The motors are prohibitively expensive, about two-thirds of our costs. We’re working with a manufacturer to build custom motors that give us both the safety and reliability at a cost that’s affordable to put into our robot. We also made the decision to work with a third-party engineering firm to help us find a way to re-engineer Jennie so she can be made at scale.
Your Kickstarter campaign in 2019 raised $60,000 and you have about 5,000 customers on a waitlist. Is the traditional investment community just as enthusiastic about your product?
My mother couldn’t remember what she had for breakfast on a given day, but she always remembered that I was working on robots for her.
There are zero venture capital firms with an investment thesis of robotic dogs that are also medical devices. It’s been a tremendous educational process of making the investment community aware of us. The vast majority of Silicon Valley investors hate hardware and the remaining ones generally steer away until you have revenue. It’s been a slog of building the network of people who love what we’re doing and who will introduce us to investors open to considering us. We made 199 presentations in 2019. The good news: We’ve found lots of angel investors who like us and we’re closing our first institutional round for $5 million now.
How has that impacted your first shipments, which were supposed to go in May 2020?
It’s taking longer to get our products out the door — we’re looking at 2021. The fundraising has caused a delay, as well as the extent of safety certifications we need to achieve because our robot will be used by children as young as 3 years old.
Your Kickstarter campaign also uncovered a huge market opportunity …
Despite 100% of our marketing being on behalf of seniors with dementia, 30% of our orders came in for other applications: children with autism, adults with major depressive disorder, PTSD. If there’s a segment of people struggling to deal with stress, our robots, which we believe reduce stress and anxiety, may be helpful. We have studies underway and more researchers continue to ask us about using our robots to study effectivity for other use cases — we want to be sure the claims we make have evidence behind them.
What’s the most rewarding thing about your work?
Talking to people who have a health issue or to health care professionals who are dealing with them. We’ve had thousands of people help us in large and small ways, telling us their stories in person, over the phone, and via social media, helping us understand how the product could be more useful.
Have you had any challenges evolving Jennie?
Philosophically, the hardest thing about product development is getting your own ego out of the way. You cling to your idea despite information that should change your point of view. The third generation of our prototype had the ability to stand and roll over. We realized that there’s not a real obvious medical benefit to that; it was just cute. But the cuteness added weight and cost to the robot, so we stripped it. The people who want this product to exist have helped us on this journey, and that’s been the best part.
Photos by Second Peninsula
CEO, Tombot, Inc.
Valencia, California, USA
MS Management, Stanford GSB, ’17
BS Political Science, UCLA, ’16
CEO, Tombot, Inc.
Executive VP, ACT Litigation Services