PEGGY JOHNSON: We put a head on, and it has cute eyes, and you can do different things with the eyes, like sparkles and hearts and things. And that actually there is a reaction with humans, a positive reaction. Robotics are kind of having a moment right now. Open jobs have doubled over the last, say, five years. And humanoids can play that role when you can’t find humans to do these jobs. This is a multi-purpose humanoid. So wherever humans go, this humanoid robot can also go.

BOB SAFIAN: That’s Peggy Johnson, CEO of Agility Robotics, one of a rising slew of companies creating robots that look and act a little like humans do. In Hollywood movies, there are plenty of fanciful stories about robots. And the prospect of AI makes that science fiction feel closer than ever. Peggy takes us past the hype, to the business opportunities that exist today for humanoid robotics — including industrial settings like Amazon warehouses, where Agility’s robot called Digit is being deployed. 

Peggy was on Rapid Response previously, after she was tapped to be CEO of augmented reality firm Magic Leap. At her new company, Agility, Peggy is navigating both a new tech market, AI’s impacts, and the tricky topic of what robots mean for human workers. I’m Bob Safian, and this is Rapid Response. 

[THEME MUSIC]

I’m Bob Safian and I’m here with Peggy Johnson, CEO of Agility Robotics. Peggy, it’s good to see you again.

JOHNSON: It’s great to see you, Bob. It’s been a while.

SAFIAN: You were last on the show in December of 2020. You’d become CEO of Magic Leap in the heat of the pandemic. Now you’ve taken on a new CEO role again at Agility. Can you catch us up on how you got from there to here?

JOHNSON: So at the time that we spoke, I think I was just a few months into the role at Magic Leap at the time. We successfully turned things around. We repointed the company toward enterprise focus. I felt like I’d done what I came there to do. I think augmented reality is still very early days. And while I love the game-changing aspect of it, and it will become part of our everyday lives at some point, the trajectory is on a slower slope, I would say, than robotics.

And when I saw the opportunity at Agility, from the outside looking at the company, it appeared they had product-market fit and a huge demand for what they could do — which is humanoid robots, putting robots into many open jobs that people don’t want. I felt like that’s the spot for me. So I made the jump, and I’m super excited about the space. Humanoids are kind of having a moment right now. 

The role that humanoid robots can play in the workforce

SAFIAN: It’s interesting how when we talked the last time, there was sort of an assumption about the acceleration of tech adoption and where that was going to lead and the change in habits. And it didn’t quite persist the way everyone predicted at that moment, right?

JOHNSON: You know, as I said, it will happen. It’s going to take some time. I think, you know, even looking at the Vision Pro, it’s going to take some time. The interesting thing about humanoids right now is they have this huge tailwind of open jobs that are kind of repetitive, dull, people don’t want. Those jobs just in the logistic areas have doubled over the last, say, five years from about 600,000 open jobs to about over a million open jobs. And humanoids can play that role there when you can’t find humans to do these jobs. 

SAFIAN: Yeah, there are all kinds of robots that are used in business, right? And a lot of them look like machines or motorized carts or things like that. But Agility has this human-like robot, which you called Digit. And one of your colleagues said robots are nothing new, but humanoid robots are. So this burst of attention on human-like robots. Why is that?

JOHNSON: What’s interesting is automation in warehouse facilities, manufacturing facilities, have been around for over a decade. And they’re magical how much automation is going on, like inside, for instance, an Amazon facility. But there’s these pockets still of humans that are maybe moving a box from a conveyor belt over to what’s called a put wall, where it’s a stacked wall that’s automated. And you can almost think of that as the Achilles heel of all of that fantastic automation. There’s these islands in the middle where humans are the glue in between, and it’s those jobs that can’t be filled. And if you can’t find the humans, humanoids are a great substitute — they can work for long hours, they don’t get hurt, they don’t get emotional, they don’t have anything that you have to worry about. It’s just: do this job — this dull, repetitive job that humans don’t typically love.

SAFIAN: When I mentioned to my wife that I was talking to someone today about robots, her first reaction was, “Oh, someone who’s taking all the jobs away.” And what you’re saying is these are jobs that —

JOHNSON: You don’t want these. 

SAFIAN: Mostly people don’t want these jobs, right? 

JOHNSON: Exactly. In fact, when they do get deployed, people feel a lot of times, the job for a human is to manage the fleet of robots. Manage the interactivity between the robots and the other automated facilities. And that is something that looks more like a career. It’s a very digital job. For instance, we have a cloud platform called Arc that helps humans manage the fleet of robots. It has interactivity with other robotics, different types of automated systems in these facilities. So the human opportunity goes from lifting boxes from A to B to running a digital platform.

SAFIAN: I’ve seen videos of Digit. Even the videos are a little spooky. Like, I don’t know, when you go to visit these facilities, does it feel natural? Like what kind of relationships do workers, the human workers, have with the robots?

JOHNSON: We thought a lot about that with Digit. For instance, we have a head-on Digit that initially we didn’t have, and it’s actually easier for a human — 

SAFIAN: Because it doesn’t really need a head to operate.  

JOHNSON: Since that time, we’ve put cameras in and other sensors that are helpful to solutions that we’re working on. And for instance, you don’t generally need five fingers. You can do many things just with a gripper hand. We can actually swap out our hands. And you can think of it as a tool. But that withstanding, we put a head on, and it has, you know, cute eyes, and you can do different things with the eyes, like sparkles and hearts and things. And that actually, there is a reaction with humans, a positive reaction that, “Oh, they’re here to help. That’s the teammate who’s doing that work I don’t like.” There is a connection there because of that head and the fact that there are eyes. I will say, these humanoids — they operate inside of a work cell, the idea of what’s called collaborative robots where you can actually interact with a human is a little bit further down the line. There’s very high safety standards. But right now we operate inside of work cells.

SAFIAN: Yeah. And I’ve seen their knees bend in a different way than humans do. Like there are things about them that are…  Is Digit a he, a she? 

JOHNSON: I refer to it as an ‘it’ oftentimes. But it’s a funny question because people do want to put characterization on the humanoid robot. We let them do what they’d like.

Using AI to develop Digit

SAFIAN: AI is changing all kinds of businesses. There was news today that OpenAI is reopening its robotics unit. I’m not sure exactly what that means. I’m not sure even they’re sure what that means. How are you using AI or thinking about using AI to change Digit?

JOHNSON: Yeah, so we’ve been using AI in the form of reinforcement learning. So it’s been part of how we’ve developed Digit. The big opportunity that we see going forward is this idea of Digit’s semantic intelligence. So giving Digit commands, unstructured. You know, the other day we said, go pick up all this trash, and Digit looks around and picks up the trash, and it puts it in the right bins — recycle bin or paper or landfill.

But going forward, how AI is going to help us is it can teach Digit new skills much more easily, rather than having an engineer program. Having the ability to tap into data built from these foundation models is going to allow us to learn new skills much more quickly. 

SAFIAN: Is it like uploading new software to upgrade the system within it? Or is the AI sort of self-learning the way generative AI is, where it learns sometimes in ways that you’re not exactly sure how it’s learning, but it’s coming up with something.

JOHNSON: Exactly. It’s the latter. And it’s why you don’t want to use AI fully to control the robot right now. ‘Cause as we know, it’s not always perfect. And when you have a robot that weighs 160 pounds or so, and has all the torque it has in its hands to lift up heavy things, you want to be controlling that with a known platform. And so, until we’re much further along with AI, we aren’t going to make that swap.

Will we see Digit in retail environments?

SAFIAN: You’re leaning into this sort of, obviously, industrial use. Is there a vision for this robot Digit to be in stores, in retail environments interacting with customers and people? Or is that sort of a different track?

JOHNSON: We’re only starting with these industrial applications because the need is so high, exasperated by the labor shortage. 

But this is a multi-purpose humanoid. So it’s meant to go off into those areas. You talked about retail, transportation. In fact, we’re already talking to customers in those spaces. I think we’re talking to five or six automotive companies. And each of them have a current need, but they have a vision for their future needs. And that for Digit will be replacing the end effectors or the hands with different sorts of tools that we can do very bespoke work for someone just by swapping out the end effector. 

You’ve seen a few of the consumer robots sort of come and go. Households, as it turns out, are pretty messy places. Things inside of a warehouse are much more structured. So the need is greater there. And that’s where we’re going to focus. And that’s where the revenue is.

SAFIAN: Every new tech has its pockets of hype, including humanoid robotics. What Peggy’s doing is tacking away from that hype into the practical, because that’s where business opportunity exists today. Of course, any new tech area has a lot of uncertainty. After the break, we’ll talk about both the good and the bad of operating amid that ambiguity, and how her experience at Microsoft working with Satya Nadella informs her calculations. Stay with us.

[AD BREAK]

Before the break, CEO of Agility Robotics, Peggy Johnson, explained how humanoid robots are filling jobs that no one else wants. Now she talks about her own job as CEO, and why she’s come to prefer working in an emerging-tech environment compared to big companies. Let’s jump back into it. 

“Staying focused on where the revenue is”

SAFIAN: So you and I have known each other for a while from Qualcomm and Microsoft, and then Magic Leap, now here. At Magic Leap, when you came in, it was a high-profile name, even if its reputation had been diminished. With Agility Robotics, you’re starting from a more modest base. So how do you think about sort of spreading the word, differentiating yourself? Or do you just feel like you stick to your knitting, and that stuff takes care of itself?

JOHNSON: Well, we are staying focused. So you’ll see us show up at supply chain conferences, automation conferences, that sort of thing. You won’t see us at some of the more generic technical conferences. We’re just staying focused on the areas where the revenue’s at, and that’s actually been powerful for us.

We’re also building an ecosystem of partners. So, for instance, we just announced a deal with Manhattan. Which is a warehouse management software company. They’ve got 1,200-1,300 customers all over the world. We can be part of their solutions and that helps a small company like us punch above our weight. 

SAFIAN: You were a first-time CEO at Magic Leap. What did you learn from that experience that you’re applying now? Like, what are you doing differently?

JOHNSON: When I first stepped in at Magic Leap, they were chasing a lot of markets and we narrowed that down to enterprise and somewhat the same here. We get a lot of questions, a lot of attention from companies who have innovation departments, for instance. I respect the innovation departments, but if they don’t have an intention to deploy and a reason and a problem to solve, we don’t focus there. We pick up and move on. We don’t want to be an interesting demo for someone’s boardroom, frankly. 

SAFIAN: You’re not just gonna be sort of experimenting, so that sometime down the road, maybe that turns into business. Like, that’s not the game you’re playing.

JOHNSON: Cause we can deploy right now, and we have an ROI for customers right now. So we want to stay, stay very, very focused. 

SAFIAN: I mean, the early stage is really different than when you’re further along as you are at Qualcomm and at Microsoft.

JOHNSON: It is different. There’s a lot of hats that you wear when you’re at an early stage. And we don’t, you know, have massive marketing teams. So we have to be creative about our marketing dollars. So it’s about working within constraints, working within a budget, quite a bit different than how I would operate at both Qualcomm and Microsoft, where you have a lot of support around you that you can tap into.

Scaling via partnerships

SAFIAN: Although you did a lot of partnership work at Microsoft also, which sounds like it’s similar to what you’re trying to put together where you are now.

JOHNSON: I definitely believe in the value of partnerships. It helps you get where you want to go much faster. You can have the vision, but I can do this all on my own. But the time it takes to get there, oftentimes trips you up and causes you to miss the opportunity altogether. So if we can find a partner that can help us in our vision, we will partner, and we’re doing that right now, talking to other… Even other robot companies.

So, Autonomous Mobile Robots. They look like a big pallet on wheels, those sorts of things. They’ll bring product up to us. We, working together with them, can unload from there and then maybe put on a conveyor belt, which is also a kind of a robot automated conveyor belt. And so those are the sorts of things that allow for a complete solution to be delivered to a customer. 

SAFIAN: So there’s an existing gap in the warehouse that Digit is filling. But if you can partner with the right people on either side, you can offer what is essentially an end-to-end solution that doesn’t currently exist.

JOHNSON: Exactly. And that’s where our customers are focused. Like they’re like, “Help me solve for this problem I have here when I can’t get people to step into these gaps.” And we have to be enterprise-hardened to do that. This is not demoware. We’re not doing backflips and making coffee. We’re working and, you know, we have to fit into their corporate IT infrastructure. We have to meet their safety standards. So all of that comes together in a complete solution because if we can’t meet that bar, we can’t enter those facilities. 

SAFIAN: They’re not going to adjust to you, though. You have to adjust to them, right? 

JOHNSON: Exactly, exactly.

Choosing start-ups over established, bigger corporations

SAFIAN: I mentioned Microsoft. When you left Microsoft, it was sort of just starting to come out of hibernation in a certain way, like it hadn’t really hit its stride. And now thanks to AI and Co-pilot and the arrangement with OpenAI, things are just cooking over there. Do you ever think about the decision to leave and wonder, like, what if I’d stayed? What if I was in the thick of all that hot AI stuff now?

JOHNSON: Actually, never. I loved my time there. I loved working for Satya Nadella. I learned so much from him. He’s clearly an iconic leader, but I remember very clearly the day I made up my mind. It was actually only a few weeks into COVID lockdown at the time. No one could see on the other side of this, and we didn’t know how long this thing was going to go on. I think nobody thought it would go on as long as it did. But I think like a lot of people, I had a little bit of time to think. My job in particular slowed down. It was biz dev, running biz dev, very outwardly facing. I got on planes all the time — that was not available anymore.

So I really sat and thought about where I was in my career, and I had always wanted to be a CEO. And I think like a lot of people, you keep saying, “well, if I just had a little more experience here, a little more experience there, then I’ll be ready.”

And I just thought, I don’t know when this thing is going to end this pandemic. I’m just going to start looking. And I did. And now that I’ve been and had the experience at Magic Leap, and now at Agility, I could never see myself going back. I like the start-up environment. I like the pace. I like the freedom, the flexibility to make decisions very, very quickly. You know, bigger, the bigger the company is, the harder it is to get to a yes or no. It’s just, “let’s debate that a little longer.” And so I love the fast-paced environment at start-ups, and I don’t see myself leaving at this point.

SAFIAN: So even a CEO role at a bigger place just isn’t as appealing right now? 

JOHNSON: It really isn’t. And you know, even in a public company, you have to do the compliance, all of those things. Those are not something I think anyone looks forward to. They’re just part of running a public company of that size. And I’m at the point in my career, I can do anything I want. And so this is what I want to do. And I absolutely love it. I just love it.

SAFIAN: Peggy, this has been great. Thank you for coming on again. It’s good to see you, 

JOHNSON: Thank you, Bob. Thanks for having me. Great to see you as well.

SAFIAN: I’m both intrigued and repelled by the idea of human-like robots. If humanoids can fill jobs nobody wants, great. Though I suspect that in time, the line may start to shift. What I totally get, 100%, is Peggy’s preference for a fast-moving start-up like work environment. Now, I’m not sure that sort of culture is dictated by size alone — even some small enterprises can be overrun with rules. But acting more quickly, taking on risks, working in a less robotic way, so to speak, that’s not only more fun but more human. It encourages the kind of applied creativity that machines aren’t equipped for. 

I hope you’ll join us next time on Rapid Response, as we dig into some quite human personal issues in the workplace — the nexus of leadership and emotion, in a high-pressured and nearly tragic situation. 

I’m Bob Safian. Thanks for listening.