Building a Startup is not for the Faint of Heart

What does it take to build a successful robotics company in today's rapidly evolving warehouse automation landscape? For OSARO, celebrating its 10-year anniversary, the answer lies in a laser focus on delivering practical solutions that create real value.

Derik Pridmore, CEO and co-founder, alongside COO Gabe Zingaretti, share the fascinating journey of navigating a decade of technological shifts and market changes while staying true to their mission. "You need to be both ambitious and aspirational, but also really practical," Pridmore explains, highlighting how OSARO has maintained that critical balance while others chased technological fads or unrealistic applications.

The robotics veterans reveal their strategic bet on piece picking—the belief that algorithms could become powerful enough to enable flexible, modular solutions across different hardware types. This focus took five years to fully develop but positioned OSARO to address multiple applications effectively while maintaining the reliability customers demand. Their approach stands in stark contrast to the industry pendulum swings: from skepticism that robotics could work effectively in warehouses to unrealistic expectations about humanoid solutions solving every challenge.

Perhaps most tellingly, the OSARO team describes their commitment to integration—working with everything from mobile robots to auto baggers, barcode scanners, and conveyor systems. This dedication makes their systems robust and valuable for customers who typically operate equipment from multiple providers. Looking ahead, they anticipate continued automation in piece picking and depalletizing, with the most exciting developments involving collaborative systems where robotic arms work with AMRs and rack storage solutions.

Whether you're a warehouse operator exploring automation options, a technology enthusiast curious about where robotics is headed, or an entrepreneur navigating your own path through hype cycles, this conversation offers valuable insights into building technology that actually works. Check out more episodes at ROSGoss.com or DwightCo.com for additional conversations at the intersection of technology and business reality.

Transcript

Introduction to Osaro Leadership

Speaker 1 0:09

Hi everyone. I'm Mandy Dwight, founder and CEO of Dwight Company, a boutique sales and marketing agency specifically for automation and robotics companies. Today I'm here. I'm also the host of this podcast, ross Goss, which I need to remember to say that every single time. So I'm here today with some industry colleagues of mine that I have great respect for, so Derek Pridmore, who is the CEO and co-founder of Osaro, and Gabe Zigaretti, who is the COO here at Osaro as well. So thank you so much for being on the podcast.

Speaker 2 0:43

Thanks for having us. Thanks for having us. Oh, of course, Welcome to our showroom in San Francisco.

Speaker 1 0:47

This is fantastic. So, you know, I came by yesterday and talked to some of your other colleagues and they gave me the tour and I was like this is awesome. I've known you, derek, since I think I met you in 2017. Like I've known you so long.

Speaker 2 1:00

Yeah.

Speaker 1 1:02

In Germany so long yeah, and in germany I was thinking about that we were in an event, loogie mad or something it was an event that a 3pl had and we had to get up there and and talk which by me, I mean my ceo at the time and then you and I sat there and just took pictures.

The Secret of 10-Year Success

Speaker 1 1:15

It was an easy day for me, but, um, lots of talks. But that was 2017. Things have you know, and I've never been here and I look at where technology is from 2017 to now. Osaro has been in business for a decade now. You've just had a big 10 year anniversary. You know how do you think about that? What's the secret to that?

Speaker 2 1:40

Uh, that's a great question. Uh, the market's moved a lot, technology's moved a lot. I think the main secret is staying focused. Staying focused on customers, staying focused on actually providing value. Yeah, there's a lot of things you can imagine doing with a robot.

Speaker 2 1:55

Um, we now live in an age when you can render anything, um and kind of propose anything, and I think you know you need to be both ambitious and kind of aspirational, but also really practical. Yeah, because warehouses need to run, they need to deliver products. You know the robots need to actually work. So I think that's always been the balance for osara. Like, just coming from my background, invested in deep mind, very much believe in, you know the potential for ai, even super intelligence, but at the same time, very aware that, um, there's hype and so you need to find middle ground. And I think it's super interesting to work in robotics because you do have to understand. You know algorithms and everything that's happening, the latest greatest cutting edge of you know whether it's llms vlas, any of that stuff, but at the same time, you're talking to customers who just you know they want to deliver toothpaste yeah and they want to do it fast and accurately and uh and so that keeps you grounded, um and and it keeps us focused.

Speaker 2 3:01

It keeps us from getting like too far field.

Speaker 3 3:03

So I think it's like that balance between aspiration and practicality. Right, yeah, yeah, yeah.

Speaker 1 3:10

No, I agree. I mean, I think the key word there is focus. You know, we have seen a lot of things change in the course of 10 years. Right, there are a lot of companies that have been very successful and I think you guys have been very like. I was excited to sit down and talk to you guys because I really do think you guys are winning and you're like no, don't say that you'll jinx us but I but I really like I, I think you guys have done a really good job. You know, with you know, we have so many customers that you have that we can't even talk about because that's how important they are you know, I mean, you guys have really, really grown and come a long way.

Speaker 1 3:44

Thank you, thank you it's.

Speaker 3 3:46

It's really focused, but it's also like trying to weed through all the noise that goes on in the industry and there is a little bit about that puppy syndrome where every time there is a new ball or new shiny light companies and especially like first-time ceos and first-time you know company owners, or they just want to chase that new thing because they think that's going to bring the new bucket of revenue investment. And really what it does, it doesn't. It just defocuses the company, burn more resources and, even though it's a cliche, it's that 80-20, right. You have to stay focused and deliver the product that ultimately works to the end and through time.

Speaker 1 4:24

Yeah, I mean yeah.

Staying Focused Amid Industry Noise

Speaker 2 4:26

That's the key working, working in production. Working in production, I mean it's funny to say but it's true there's, especially in the last five years. This is a super interesting market to have been in for 10 years, because I would say from 2015 to 2019, it was like this can't possibly work. And then, of course, there was COVID, and then somehow now it's like, well, but there'll be a humanoid that parachutes in from Mars and it'll do everything right and it's like no kind of neither one of those was right. It's like it was always going to work and it's going to be modular.

Speaker 2 5:03

Our philosophy here is modular, flexible robotics, tight integrations, high uptime, high ROI Keeps us from having to build the hardware. We manage the algorithms, we manage the fleet and our customers get kind of the best of. You get the robot that fits the use case with its depalatizing piece picking, you get the high uptime and you don't need to worry about essentially consuming someone's science project. And what gets especially tricky, as Gabe was pointing out, is the pressure comes from both sides, because you can have customers who you know they legitimately want something in production, but they also don't know what's possible and they're making a lot of asks. And then on the flip side. We're in Silicon Valley, we're VC backed Um. Vcs can also put that pressure on you. So you know you can be scaling, uh, an application in production with customers at multiple sites. And then someone starts talking about LLMs and you know they want to know how your robots work with LLMs and it's like that's a complicated question, not the most important one, it's not even like 10th on the list, yeah.

Speaker 2 6:08

You know we can get into it.

Speaker 3 6:10

It also like doesn't have to work with LLM, right? So there is like wish and necessity, and I think there is not a confusion. But they're trying to blend the peak of the peak of technology with what you really need, right? The peak of the peak of technology, what you really need, right? So if you want to go to work, you can do it in a Prius, you don't need to drive a Lamborghini right, you can.

Speaker 3 6:29

If you want to, yeah, you could, it's probably not necessarily what you want to do on every day, right? So that is what do you really need and what you want and what you want. If it's not, what you need is what you need is going to come at a price. And now the ROI is breaking down, right, you can't make it work anymore, right? So what we do really well is a process of education and collaboration with our customers, which is not just listening to what they want, but also like bringing down to the reality of the ROI, which ultimately, like the numbers, have to work, right.

Speaker 1 6:59

Yeah, no, I agree, I mean customers. They need the value. I mean tech specs, all that thing, it doesn't matter if it's not working on the factory floor. Uptime is really everything.

Speaker 2 7:09

Yep, people work really well. You know they sometimes make mistakes, but more or less you can rely on them and they can handle exceptions.

Speaker 2 7:20

That's the thing that they're really good at right, and so it's one thing to go put a robot in there for a short period of time, but to actually in our case, each robot replaces two people. So to pull two people out and have a robot working, you know, 18 hours a day, you need not just the algorithms but you need a lot of other things, and that's what we've spent a ton of time over the last decade building sort of fleet management tools, all the integrations so that we can, so that we can integrate, so that we can redeploy modular robots and be flexible enough to support customers' needs, but always do it in a way that's practical, so that you know we're staging through. I mean, one of the questions you were asking earlier was you know, how did we, how did we choose, kind of where to target in the technology stack, like how much to do versus how easy to make the problem? So, specifically the warehouse market, for instance, you could think parcel induction, you can think palletizing, depalletizing, and then you can do piece picking and you can get into combining that. You know how do you combine it with rack storage systems, conveyor belts, amrs.

Core Technology and Modular Approach

Speaker 2 8:29

For us, what we thought about was what's the core algorithmic technology that we need to build such that we can actually tackle all of these use cases. And, in particular, you know what we I think the decision we made that was a little bit out there in 2015, but turned out to be right on target was that we believed piece picking was possible. We believed the algorithms were powerful enough that you could do it in a modular, flexible way so you could do cross hardware. So if you look back here, you're going to see probably eight different types of robots in the background. They're all run by the same core control technology and the same core perception algorithms, but they're fine-tuned and reconfigured so that if you need to pick up a 50-pound box, use a big robot. If you need to pick up something very small like watch batteries, use a very small robot.

Speaker 2 9:19

That was the hard bet that we made in the beginning and spent a lot of time. You know took us five years of sort of technology development and piloting with customers to build out all the integrations. But once you have that, then you're in a position to do all those applications well meaning. In theory, old style technologies could do depilotizing, but in practice they often fall over because there's something on the surface whether it's damage or just idiosyncratic box markings that make it difficult. And so if you built your system because you knew you were going to be picking you know a multicolored foil, you know foil reflective Colgate box then you're not worried when someone's boxes have sort of wild printing on them. And so that was the key, the vision system.

Speaker 2 10:10

And then from there sort of spinning the other way around and getting really pragmatic and be like, okay, no more, we're just going to focus on integrations now and scaling the fleet as opposed to, for instance, can we put this on a toy dog or can we put this on a humanoid. That's kind of where we drew the line, because we want it to be very practical and useful for our customers, and that's still where we're focused. You talked about scaling with customers. Now we're at the point where we've got, you know, deployments with customers at multiple sites where they actually want more, and then you get into the part of the problem where you can just kind of clone and scale up, which is a better, more fun place to be.

Speaker 1 10:50

Well, land and expand. I mean my salesperson heart just loves that you know, because that's more robots deployed, better ROI for the customer as well.

Speaker 2 10:58

Yeah, when there's multiple, Exactly they don't want, they would rather de-risk something, have it provide value and then scale it up and then do the next thing. Not de-risk something, have it provide value and then scale it up and then do the next thing. Not de-risk something, and then go de-risk another thing, and that's yeah. I think that's made sense for our customers. It's made sense for us, yeah.

Integration Challenges and Solutions

Speaker 1 11:18

Yeah, and I think, like walking around here yesterday, I was so impressed with the amount of the equipment that you've integrated with you know, and I think that tells the story, because people don't realize that a piece of technology just doesn't function on its own. It is part of a solution. So I saw mobile robots here, I saw auto baggers, I saw every piece of equipment that a customer has right now in their warehouse.

Speaker 2 11:42

Yeah, especially when you start looking at piece picking whether it's a barcode scan or some other sort of value added behavior. Like you know, we've had customers who want to put coupons and things like that. So car dispensers, barcode scanners, auto baggers you mentioned and then you get into conveyance so conveyors, amrs. Beyond the piece picking problem, the manipulation problem, there's this integration problem. It's really important and it's actually a great opportunity to not just make the system more robust but provide more value to customers.

Speaker 2 12:19

So, like I said, once we had the core technology working and that became a big focus for us was just expanding the set of integrations that we have to make sure that we work with. Kind of everything can play with everything and then also API-ifying, modularizing it so that when we work with integrator partners because again, it's a complex ecosystem you know big customers are generally consuming warehouses or solutions from multiple providers, so you need to play well with other people's equipment and I think that's been a strength for us. It's something we're not afraid of. Pretty much every one of our deployments has some kind of there is always something.

Speaker 3 12:56

This recently was the first time that we deployed robotics out of bagging with consolidation. So usually you get a toad and there is, like you know, other things you need that you got to put them a in a in a poly bag, right in another bag, or, uh, this actually the, the toad that comes as multiple SKUs for different customers, so you have to sort them and then go back and repeat them so you can consolidate the order and sold out robotically and, uh, it's the first time that we've done it for sure, but it's the first time that I've done it for sure, but it's the first time that I think I've seen it just in general. So you know, it's not like. You know, everybody can claim to have AI and everybody can say, oh, I have Peacepeak in, but like, do you have Peacepeak in and do you have it all the way through, where it is a product that a customer can adjust and use and pay for it for the foreseeable future, right, and I think that's where we're not ahead of everybody else's.

Speaker 3 13:49

Well, you stayed the course. Yeah, I stayed the course.

Speaker 1 13:53

And it's noisy. I mean it is noisy. We have seen a lot of things come and go in the past 10 years. A lot of companies come and go, there's been a lot of changes and sometimes I feel like things are going around.

Speaker 2 14:13

Generally try to ignore it. Obviously, you need to pay attention to what your customers are saying, and you sort of have to pay attention to what VCs are saying if they're paying attention, but otherwise we try to stay focused on what makes sense in the real world.

Speaker 2 14:32

I've told some of our backers this and we talked about it to our customers. The nice thing about doing AI in the context of robotics is, at the end of the day, there's actually a physical operation that's happening and you can use that to massively constrain what's possible and what's efficient. So you can imagine all kinds of crazy things. You could juggle these items. You could build a robot that does that. But then you ask yourself, is that useful? And I know that's a ridiculous example, but it turns out many of the other things that you see and have seen over the last 10 years were also ridiculous. So you know, if you just say, is there a simpler way? You know constantly, even if it means it doesn't seem cool, it turns out there's enough hard stuff to do. You don't actually have to make anything harder than it needs to be. If you just relentlessly try to make it simple, it'll still be too hard, and I think that's mostly how we've tried to avoid the noise.

Speaker 3 15:29

Also say no to things.

Speaker 1 15:31

It's important to say no.

Future Trends in Warehouse Robotics

Speaker 3 15:33

It is important to say no, a lot of companies can't, because it gets harder. You to say no, yeah, it is important to say no, a lot of companies can't right, yes, because it gets harder. You know you develop something and it doesn't quite work. And it's not quite. You know you're 99.5 percent of time your throughput is not there, and that's where you know you're looking at something else. You're like well, but if we do that, that's gonna work. And then you do the same thing. Right, you get that 80, it's there, and then it doesn't quite work. You keep on moving. So we say no, we are not going to course deviate, we're going to course correct. Right, because we're learning things. But we're on that trajectory. We're going north and we're going to stay north. Right, the company and the team, to their credit. They really believed it and they're marching in that direction. Yeah, and they're marching in that direction.

Speaker 1 16:14

Yeah, and as far as that goes, like managing the noise, but you know people like you, you see everything that's out there. You know in different ways, right. What's next? What are the next trends?

Speaker 2 16:25

In warehouse specifically, or robotics.

Speaker 1 16:28

Warehouse robotics Like what are you seeing that you're like, wow, that might be a thing or that might not be a thing.

Speaker 2 16:35

There's a lot going on, might be a thing or that might not be a thing. There's a lot going on, um, and it's definitely going to remain an ecosystem, meaning um, there's probably gonna be five answers to your question and maybe everyone would say different things depending on what they're seeing like. I, for instance, continue to see peace picking, uh, depolitizingetizing. That's all going to get automated, or at least 80, 90 of it, I think. Um, for our money, arms that are um collaborating with amrs or rec storage systems so that every component in the system is moving, essentially 100 of the. That's kind of where we see things headed.

Speaker 2 17:16

So and I'm just saying that to contrast it with not a big believer in kind of a one size fits all, whether it's a humanoid or like a quadruped that's going to be able to do many different tasks, partially because of the utilization problem, partially because of that value-added kind of necessity that I mentioned, where you need to stay integrated with a lot of things. Um, like humanoids would probably have to wind up doing a barcode scan and typing and a bunch of other stuff and it turns out you just do that all electronically. So the the trend I see is just um, probably, I think definitely the next that's happening now but in the next two, three years, having kind of the arms, come up against the AMRs and work with them as well as the rack storage systems, I think that's interesting, it's valuable and you know we're doing that. We see customers who want it. So for my money, I think that's probably one of the more interesting things to watch. So for my money, I think that's probably one of the more interesting things to watch.

Speaker 3 18:17

It's definitely like the most efficient way of looking at warehouse automation with ROI, right? So an AMR is supposed to just move, so that's what they do, right? So you're paying for something that moves, and the robotic arm is supposed to move as well, but it's in a fixed position and move things, right. So the combination of the two is almost like you're getting the best of both worlds the part that is supposed to walk or roll or stroll is always doing that, and the part that is supposed to move and pick is always doing that, right yeah no, I agree.

Speaker 1 18:47

I mean that's tangible right. That's what people need right now. You know, labor challenges are always going to be there.

Speaker 3 18:58

A lot of brownfields, you know, and you guys do a lot of brownfields, yeah, yeah, I mean, I mean I don't come from what, from the material handling space. And when I walked in the first warehouse, I mean these things are massive and there are millions and millions of dollars of equipment in it. They are not going to be dismantled anytime soon. Right, you have to retrofit. Yeah, and building a new warehouse is like two to three years ahead, right, and in two to three years the technology is moving. So as you build the warehouse, you got to see whether the technology is moving because you can accommodate it. Otherwise you'd be three years behind on something that is brand new.

Closing Thoughts and Celebration

Speaker 1 19:27

Yeah, absolutely so. You know we're coming up on time and I know that you have an all hands upstairs in a little bit. So one thing I did bring for you guys you know I talked at the beginning. You know I've known you guys for a long time. We sold against each other for a while you know, so, yeah, yeah, yeah, we forgive you.

Speaker 2 19:46

Yeah, I know.

Speaker 1 19:48

So I sold a competing product in 2017. I was like, oh my God, that Osaro. I'm so nervous about them and you know you guys were tough, to sell against. You won. Like you, I mean the product slightly different market, but yeah, yeah, slightly different, but you know, you guys did a really good job.

Speaker 3 20:04

You sold, so I saw this online oh oh, wow you guys, thank you so I wanted to bring this to you guys, so this is for Osaro.

Speaker 2 20:14

Congratulations, you got me so yeah. This reminds me of the More you Know star.

Speaker 1 20:22

Maybe we'll close the podcast with that.

Speaker 3 20:27

Thanks for coming out.

Speaker 2 20:29

Thank you for the trophy. Thank you so much.

Speaker 3 20:32

Again.

Speaker 1 20:33

Mandy, dwight, dwightcocom. Anything that you see it's on RossGosscom Dwightcocom. Anything that you see it's on RossGosscom, dwightcocom. And thank you guys so much for sitting with me.

Speaker 2 20:41

This was really fun. Thanks, guys.