Episode 14: Robotics That Work: How to Pilot, Prove, and Scale with Sam Abidi

Robotics is moving from the lab to the loading dock, farm, and job site—and the playbook is changing. In this episode, Sam Abidi—robotics expert and member of the U-Path Advisory team—explains how to pick the right first use cases, redesign operations around machines, and structure partnerships that actually scale. We cover ROI beyond simple labor displacement, how CVCs de-risk technical bets, and why the integration playbook for AI agents mirrors what’s worked in robotics.

What we cover

• Real deployments gaining traction: sidewalk delivery robots, trailer unloaders, and autonomous construction equipment
• The integration playbook: select use cases, re-engineer operations, define KPIs, and stage-gate rollouts
• ROI that compounds: productivity and quality gains vs. pure labor savings
• Partnering to scale: robotics company ↔ industrial customer, plus OEM/JV dynamics
• Build, buy, or partner? When M&A “programs-in-a-box” make sense
• Workforce strategy: reskilling to remote ops and advanced maintenance
• CVC’s edge: technical diligence, creative structures, and internal championing
• Why AI agent adoption will follow the robotics integration model

Guest: Sam Abidi, Robotics & Autonomy Advisor, U-Path Advisory; former CCO at Embark Trucks; founder, Apex Advisors.

Subscribe to Inside CVC and explore more episodes at u-path.com/podcast.

 The CVC Open Innovation Summit Europe takes place November 25–26 in Berlin, during AsiaBerlin Innovation Week. Positioned as a bridge between Europe, the U.S., and Asia’s fast-growing ecosystems, the summit is a curated, invitation-only forum for senior leaders to tackle the issues shaping innovation: AI, geopolitics, energy resilience, health, sustainability, and more. Listeners are invited to request an invitation at cvc-summit.com → E

Support the show

Catch up on all episodes of Inside CVC at www.u-path.com/podcast.

Transcript

Steve: 0:00

Thank you. Welcome to Inside CVC, the podcast that brings together leaders in innovation and capital investment to explore the trends shaping the business of corporate venture capital. I'm your host, Steve Smith, and together with Philip Willigman, we're speaking to corporate investors, entrepreneurs, and ecosystem builders driving the future of innovation. Each episode, we dive into the strategies, partnerships, and big ideas behind venture investing at the intersection of business growth and emerging technology. Inside CVC is brought to you by U-Path Advisors, helping corporations and startups unlock sustainable growth through strategic partnerships. To learn more, visit upath.com. That's the letter U-Path.com. And to catch up on all of our episodes, search Inside CVC on your favorite podcast platform or visit upath.com forward slash podcast. In this episode, we're joined by Sam Abidi, an expert in robotics and a member of the UPath advisory team. Sam has spent his career at the intersection of robotics, autonomy, and corporate partnerships, helping companies navigate the complexity of integrating machines into real-world operations. We talk about the breakthroughs that are moving from labs into warehouses, farms, and construction sites, the challenges of scaling robotics inside large organizations, and the role of corporate venture capital in making the right bets today to prepare for tomorrow. Here's our conversation with Sam Abidi. Sam, welcome to Inside CVC. Thanks so much for joining us on the show today. How are you? I'm

Sam: 1:48

doing great. Thanks for having me on the show.

Steve: 1:50

Absolutely. Thank you for taking some time out of your schedule. Really, really excited about today's conversation about robotics. Such an interesting time when you think about advancements, not only in the technology and the materials but also the software, the AI, the coding that's running these machines. Why don't we start with sort of your point of view on where robotics is today? What are some of the use cases you're seeing across industries like logistics, manufacturing, agriculture that are really exciting you?

Sam: 2:21

Yeah, so I think you could take any industry today and you could give three, four, five examples of breakthroughs that are happening on the robotics front. Maybe I'll just hit on a few that I think are really hitting their stride in commercialization. There's obviously a ton that are getting a lot of coverage in the press, but what I find more interesting is those that are kind of tipping over from idea to actual value add in the environments that they're deployed. I'll just hit on a few of those. I think one that's really exciting is in delivery. You're seeing a lot of what they call sidewalk robots. These are essentially coolers on wheels that are taking food from restaurants to people's homes. This is a segment that's kind of breaking out from the pack, which might be something like a Waymo delivering your food or a drone delivering your food. And you're seeing a lot of traction today on things like college campuses. So from the dining hall to people's homes. And it's one of those examples where you got the right use case in robotics. It's slow, so it's safer. It's on the sidewalks. It's a more controlled environment. It's a relatively low cost platform. And it's solving a very pertinent need, food delivery, which is existing already. And so that's kind of when I think about like real man, that's great robotic examples. That's probably one of the ones that comes top of mind in the food delivery space. I think in logistics, inside of the warehouse, you're seeing a really interesting use case pick up with loading and unloading of trailers. So everyone here probably has a general idea of how a warehouse works, but you've effectively got trucks coming in on one side, they're getting unloaded, they're getting all their materials getting sorted, and then it's getting sent somewhere on the other side. The unloading of the truck is actually a bottleneck. right now in warehouse operations, and you've seen a couple of robotic companies pop up in the last few years that are sending a robot into the trailer and grabbing all the boxes and loading them onto a conveyor belt so that multiple people can attack the packages and put them on pallets and send them into the warehouse. And this is one of those examples where it's not some very complex task that the robot's doing, but it's better suited to essentially alleviate the bottleneck than a few people inside of the trailer. And it's not just labor displacement here. It's actually freeing up one of the critical bottlenecks. And so the value is much greater than just simply displacing a person. So this is just another example and another vertical. A last one I'll give you is construction. You've seen in the last year or two years, multiple companies pop up that are tackling automation of construction equipment, things like bulldozers and excavators. And for maybe five, six years, this was a pretty dormant space, maybe one or two players. And in the last year, it's tripled or quadrupled in the number of folks tackling it. And I think it's because you're watching a lot of the on-road autonomy teams move into this easier use case, right? It's a somewhat controlled environment. You can dictate the pace you do the work at. And so this one excites me because it's an example where the great efforts of the on-road autonomy world are being leveraged to essentially solve a use case that's a little bit more near term, a little bit more solvable. So those are, as I kind of step back and look at the hundred different robotic use cases that are taking off today, these are a few that kind of really excite me for slightly different reasons, but all because they're starting to hit their stride and actually break out into delivering value to the customer.

Steve: 5:55

Your first use case, the robotic delivery of groceries, resonated with me in particular for two reasons. I think when people, at least today in the United States, see some of those things, I think the natural assumption is those use cases are confined to these very urban, very tech, right? Whether it's the Austins, whether it's the San Francisco's of the world, etc. Over the fall and winter, I had the opportunity to be in two very interesting places, San Francisco, where I ran into one of these robots and it was recorded. It was really cool. But also on a more personal thing, my friends and I, we drive here from St. Louis to Bentonville, Arkansas every year to do a mountain biking trip. And on the way down, we stopped in Rolla, Missouri, home of Missouri S&T. And we actually almost ran into over one of these robots, right? And so in terms of the unexpectancy of seeing this technology scale in the middle of Missouri, where you would think of these things are being reserved or at least scaling earlier on one of the two coasts or in these more progressive urban cities. To me, I just think, as I hear you say that, it is reflective of, to me, of how much these sorts of very real value-oriented use cases are already scaling in the United States, certainly globally as well. When you talk about these use cases, when you talk about these early examples of delivering value and scaling, what do you think are some of the tailwinds and the headwinds in today's age that are helping catalyze that sort of growth and scaling.

Sam: 7:36

Yeah. So I'll start maybe on a tailwind. So I think this is probably one of the biggest things that's happened in the robotics space in the last 24, 36 months is you've had just a step function change in the way you get synthetic data and run simulation. And I'll explain a little bit about robotics to help contextualize why that's important. So in software, at the end of every day at a thousands of tests. Unit tests, that's test small pieces of the code. There's workflow tests that test entire operations. And you run thousands of them. And it's the equivalent of hundreds of thousands of user hours. In robotics, that testing has to be done often in the physical world. And so it's a thousand times or a million times more expensive to test in the physical world than it is in software. And so robotics has long been trying to solve that by doing more of the testing and simulation. But up until a few years years ago, you could only generate synthetic data of such quality and you could only generate kind of edge cases or unique things that would happen to the robot in the physical world. You could only generate so many of those in the digital world. And the generative AI breakthroughs that have happened in the last two years have allowed you to just multiply that a hundredfold. Synthetic data can be created at a much higher fidelity and probably more importantly, scenarios can be created at a much higher fidelity. And this has allowed robotics companies to move a lot more of their physical world testing into the soft world world, which reduces costs significantly and increases cycle time. And what do those two things really translate to? They translate to more robotic use cases making sense. Today, there's a lot of robotic use cases you can look at and you can say there's not enough data and there's not enough physical training, so we shouldn't approach it. More use cases have moved out of that bucket into the we should approach this because of the advances in synthetic data and testing. So I think that's hands down, in my opinion, the most exciting tailwind for the robotics space. And then on the headwind side, I'd say that the number one thing that continues to kind of worry me is hardware costs. We're seeing a lot of great stuff out of other areas of the world, a lot of them used out of China of relatively low cost humanoids or quadruped robots or even AMRs. But when you really get to five nines of reliability robotic hardware, the prices are still extremely high. And that limits the number of use cases that make sense. You have to have this incredibly high utilization of the asset, multiple shifts. It has to be doing something that's both labor displacement and incremental value. And so the headwind that I'm really excited to see get kind of beaten down is these hardware costs. sensors, actuators. Teams have iterated to bring these down by a factor of 10. I really think they have to go down by another factor of 10 to really open up another frontier for robotics. So those would be kind of my tailwind, headwind at this moment in

Philipp: 10:42

time. Sam, it's great to have you on the show. The three of us all have worked in autonomy for quite some time in our Deloitte days, but you have started working with robotics back when you were a teenager in your parents' robotics lab. So I feel like there couldn't be any better person on the call today. Can you help our audience actually understand, you talked about the advances and what robotics can do, but what does it really take to integrate robots into company operations? It's obviously not a usual thing that suddenly a robot is walking around, and I just saw it last week in Paris, a human robot taking care of carrying tires from A to B. It's slightly different than if you work in a human environment. So what does it take? And also, what are the different factors you see robotics companies and corporates having to consider to really make it happen?

Sam: 11:32

It's a great question, Philip. I think as much as the technical development is an incredibly tough endeavor in robotics, I really think it's half the question and the other half is operational and how you get it integrated into the business. So I'm happy to kind of talk through the different factors that I think are important to consider. So anytime you have a robotics company approaching a large corporate with a proposal to integrate that robot into their business, there's typically, one, an effort to figure out what the best use case is. Even if it's something as simple as one of these loading-unloading robots that I was talking about earlier that go into a truck and unload the boxes, you want to quickly ask for that warehouse which trucks come with boxes that are of the type this robot can unload, that are of the right size, that are configured neatly on pallets, and which trucks not that way. So let's segment out where we have this robot serve to be an area where it's well suited. Then once you've got that use case identified, you move into this question of how does the operating model need to be updated? I think the number one reason robotics programs in large corporates kind of die on the vine is because the company and the robotics player don't take the time to reinvent the operation. It is rare that the The way humans do it is the way robots are best situated to do it. So you take something like this loading unloading, you now might, for example, need to put lighting in the truck so the robot can better see the boxes. And so there might be actually a step where a human steps in and puts two lights up and then the robot goes. They never did that for the human who walked into the trailer before, but they might need to for the robot. And I think taking the time to accommodate the robot is one of the most missed steps. And this can get really complicated if you think about how you need to accommodate for a robot on a construction site or a mine site or an AMR robot that's running around a warehouse. That can become a work stream in and of itself. Once that operation has been redefined, there's an effort around defining the business model and the business case. Very often folks will just walk into a pilot and say, well, if it can do the job the human did, we're happy. But you need to be much more more prescriptive in the KPIs because it's rare that it's going to do the job the same way. It might do it slower, but with higher quality. It might do it faster, but the quality is lower and so humans have to intervene. And you have to really dissect that business case so you know exactly what you're measuring because otherwise you'll get to the end of it and someone will say, well, it did it slower. We shouldn't move on. But if you actually were tracking the quality, you'd see that the downstream quality benefits outweighed. And then the last thing you have to do, and this is all before you get going, is develop the rollout plan. So I mentioned at the very beginning this notion of you might point it at the trucks that have goods that are well-situated, but you don't want to end there. You want to eventually expand to the trucks that have boxes that are not well-situated. That requires incremental development of the system, redoing of the operating processes for those trucks, and then there's probably a next segment and a next segment. And if you don't chart out that network-wide or business-wide rollout. At the very beginning, you will never be making the right decisions at the beginning to get to the second stage. It's why we set strategies in any part of our business. It's so we know what's coming down the pipe so that we can be making decisions today to set us up well for that. And I watch this get missed a lot in robotics because someone in the org just wants to get a pilot going. And I think that that's often a way to kind of set it off on the wrong foot. So I think it's important to do all these things that I've mentioned up front before you get going, because it'll be well worth kind of the chewing glass exercise that you've got to do with the teams when you get down the

Philipp: 15:36

road. So on that, I mean, obviously in organizations, when change comes, there's a lot of naysayers, right? How do you suggest like robotics companies and the CEOs and their sales teams really work with the corporate naysayers and convince them, oh yeah, come in, do this. We'll just do a quick POC. Don't really care about the outcome. Just somebody told them you have to test it. But how do you really help the robotics companies to set themselves up for success? Because I've seen this in my world where you have a conversation with a corporate and say, oh yeah, we're going to test every human robot on the planet right now. So you just come in and we just do the test with you as well, right? But they don't really share all the details and all the characteristics. So the only thing they can is fail, right? Even though they might have the best solution, who knows? But how do you overcome that? Because I've seen a lot of corporate leaders will say, okay, well, I don't really want that anyway, but just let's do it.

Sam: 18:03

A lot of this comes back to that operating model I was talking about. So when you chart out the life cycle of what the robot's going to touch, both the things you have to do before it's interacting with your business and the way that those physical goods get used downstream, you end up with a lot of different parties in the org being involved. Three, four, five different business units could be touched. And you effectively need to take the time to go to each of them and explain how this negatively and positively affects them. I think one of the disservices robotics companies will do to themselves is they will just tout the benefit. But I think it's really important to be frank with the different stakeholders on here is the give to get. Here is where you're going to have to adjust the way you work. Palettes are going to come in in a sporadic manner. You are going to have to accept things in the evening. But what you're going to get for it is higher quality, better throughput, etc. And essentially walking each party through their specific trade-off up front has been an exercise that takes a tremendous amount of time, but you almost certainly break down when you don't take the time to do it. And that means, frankly, putting together five different business cases for five different business units who are all going to be affected by this. Like on a construction site, the general contractor, they have a labor and workforce change that's going to occur when robots come on site. And so they have to train people up and train people down. The owner of the site is going to receive faster completion of the construction project, but they're going to likely incur a higher charge for the autonomous technology. The subs on the site are going to have to give up some of their service But they're going to get a more seamless interaction with the robots in the areas where they used to interact with humans who were sporadic in their behavior. Each one of those parties has to be sold separately on what is a rather complex business case. And so that's kind of the process you have to undergo. It's no different than if you're doing a digital transformation, right? We were at Deloitte. We've seen dozens of these giant ERP implementations, right? You go to accounting, you go to finance, you go to HR. and you spend days selling each of them. Robotics is the physical manifestation of that, and the same process applies. It just tends to be a little more physical in nature, which makes it more interesting. It's not as easy as explaining your workflow at your desk changes. Where you show up in the morning might actually change.

Steve: 20:42

Sam, what you describe, as we all know, is hard enough to do in your own organization, right, man? But I want to go, you stick with your robotic warehouse. use case. I believe in the very near future, you are going to see this customer experience where I send my robo-enabled autonomous vehicle to go pick up groceries, and that is actually my vehicle showing up in a warehouse, the hood popping, my order being dropped in my trunk by one of these automated forklifts or whatever, however you want to call it. And in real terms, that's a grocery store, a logistics company, an OEM, all working together in harmony. And so as you think about these cross-brand, these cross-industry partnerships, how do you have the conversation that you just described to create buy-in not only in my organization, but how do you have that conversation when you're talking about a number of different organizations working together? to invest in robotics. And what role does the robotics company itself have in bringing that ecosystem of companies together?

Sam: 22:02

This is a great question. I mean, this is where I spend most of my time with clients. So I typically work with the robotics companies and it's helping bring these ecosystems together and structuring those partnerships. As far as who has to own it, it is mostly the robotics company. They are the one who have the economic incentive to bring this all together. And so they end up having to carry the torch. I think about it in kind of two classes. You've got one, what is the robotic to customer relationship look like? And then you have all the other partnerships. So I'll talk first about the robotic to customer relationship. A lot of robotics companies might hire a salesperson who is looking to essentially sell this robot forward, similar to how they would have a software, which is typically an arm's length transaction of here's my software. Here's some support integrated and then you're off to the races. And I think anything in the physical world does not work that easily. And so I kind of tout that the relationship between the robotics company and the industrial company is going to run a spectrum from a partnership all the way up to a joint venture. And on the partnership side, which is where I try to steer 80% of deals, both parties are putting in resources. You are likely requesting a team of people at the industrial who's going to help with all those things I was mentioning earlier, building the operations plan, finding the use cases, the business plan, et cetera, running the rollout, and an equally sized team from the robotics company. You might be putting capital in to co-develop the robotic solution. It's very often a robotic solution gets into an industrial setting, and it actually needs incremental development to work in that company's domain. And so you're structuring this partnership where you're both investing at the to develop a strategy together. And then you're essentially stage gating your rollout. You're running your first pilot against that well-defined business case. If you meet it, you're crossing that gate and you're possibly going to a regional rollout. And then if you meet that, you're possibly going to a nationwide rollout. And then you're likely going to a subsequent phase, which is filling in the gaps, those really hard sites or use cases that weren't handled. This is often a multi-year journey and it requires commitments from both sides, right? Because you're both going to drop a bunch of resources in in the first year, and you both want to get something out of that. The robotics company wants some guarantees of volume and demand, and the industrial likely wants some guarantees of favorable support and pricing. And so all those things have to come together into a partnership agreement. And you'd hope to structure that as something that you can, with 80% accuracy, take from industrial customer to industrial customer. If I'm doing construction automation, I'm going from general contractor to general contractor trying to structure that deal. If I'm doing it in a warehouse, I'm going from logistics company to logistics company trying to structure that. And that handles my customer to robotics company relationship. There's still all the other ecosystem relationships. So you mentioned OEM. This is typically your next biggest relationship that you've got to get in place. You often find yourself having to essentially put a joint venture together with your OEM partners, right? Because they're putting in a tremendous amount of development and brand risk to build up your vision. Often you have a prototype of what your robot's going to look like or a minimum batch run of let's say a hundred, but they're the ones responsible of getting you a thousand units. You might be going on top of their platform. And so there's a whole host of economics to work out, supply chain to work out, volume ramps that are typically against these customer relationships. And so that tends to take a different flavor that looks a little bit more like a joint venture because it's going to be one or two or three OEM relationships that are going to allow you to scale your units across 10, 20, 30 customers. So those are the two most critical. Obviously, I say that the customer one, an order of magnitude more important because it's where your revenue is going to come from. And those are a little bit of guiding points on how I've kind of helped structure those in the past.

Steve: 26:21

As you describe this, 100 comes to mind as an example and I'm not not very keen on the merits of the actual transaction but I think about Hyundai and I think about the role that they have in the movement of goods through shipping the role that they have in terms of heavy equipment they build bulldozers they build right obviously they build cars they have a very clear view on the future of urban air mobility and they acquired Boston dynamics probably one of the best i don't most known companies most familiar company when it comes to robots and the types of robots that they produce again i'm not up to speed on what the transaction was but if i'm if i'm in hyundai i'm curious what role does cvc's what role do cvc's have in creating that demand there was somebody i have to believe within hyundai that said at one point we ought to go buy boston dynamics And this is why, right? And so that role of CVC, you talk about the importance of robotics company. What's the role of the CVC organization into driving these partnerships and driving this investment and scaling of innovation?

Sam: 27:36

Yeah. And look, I think when you step back and you look at where CVCs tend to exist in kind of the Fortune 500, it's a little bit more skewed towards these large industrials. And I think there's a reason for that. I think their businesses are so complex and so multifaceted. that you need an entire org running the interface with innovation. And the CVCs have kind of emerged to serve that need inside of large industrials. And I think one of the best use cases for them to sink their teeth in and add value is robotics. And I can talk a little bit about why. It goes back to some of the points I was making earlier about the number of parties you have to get aligned and the length of time that a robotics company and industrial have to work together before value gets created, right? All those things I was talking about earlier have to happen before you actually start to both see the fruits of your labor. And CVCs have an ability to help in a few ways, right? They can essentially help tell that value prop story to the cross-functional players within their organization who need to be brought on side. That's one huge area that I've seen CVCs help in deals that I've been involved in. They've been the ones to go speak the language of said industrial and help explain the business case that we, the robotics company, put together. So I think that's a huge one. The second one, which is more traditionally how we would see them helping, is in creative financial structure. Robotics deals, I think, have some of the more interesting financial elements to them of any type of deal. It's very common. There's obviously an investment into the robotics company, which the CVC is right down their lane. But there's also often these creative warrant agreements or upside for the CVC and the industrial in cases where the robotic company really has a breakout. And the reason I think these emerge is the industrial is often providing tremendous help to the robotics company by providing them an environment to test and develop in. And that creates an opening for the CVC to structure really interesting win-win deals on whether it's warrants or options, types of to revenue targets both within their specific company or the broader segment or just the industry generally. And so I think those are two areas where I've seen CVCs kind of grease the skids on robotic industrial deals in a really good way. I've had many deals that have kind of gotten across the line because of the CVC's creativity, support, and just legwork.

Philipp: 30:16

I mean, that's obviously an area we are focusing on within U-Pass. And I also feel like CVC extremely helpful to help navigate the robotics companies right within these massive organizations and being a good sounding board to really make sure that some of these proof of concepts come to fruition and it's kind of like going to the next I think very important topic which is like the return on investment it's something I had in my conversation with corporates there's like the one side where I think they're realistic to say they know how long it is actually sometimes taking and then I had one conversation lately and would love to hear your thoughts, Them, how you're dealing with this, where they were like, yeah, we would accept only like return on investment after like 10 to 12 months. If there's no return, then we don't even start doing this. And we were even kind of like then talking a little bit deeper about this and said, are you really considering everything into a return on investment, which you have to? Because like, I mean, obviously, if you, for example, replace a human with a robotics who may have done a job which was like extremely physical, complicated, or may even have led to that this human could potentially only do this job for like 10 to 15 months. So do you really think about it in the right way? So what is your thoughts on that? How do you think about measuring return of investment? What are really the right KPIs when it comes to robotics? And what is realistic to have a fair assessment from the startup's perspective, but also from the corporate's?

Sam: 31:52

Yeah. This is a great question. I'll talk. So maybe I'll break it up into what are the sources of ROI, like what helps generate the ROI and then timeline and timeline will kind of flow from the sources. So a lot of times people will think first principles of robotics, the savings is in labor displacement. And I think the first savings often is in labor displacement. But I think a robotics venture that is purely labor savings has a very low chance of succeeding. You need something incremental of that, and that typically comes in the form of increased productivity. This robot does the work of two or three or five humans, or increasing quality. Those values end up being multiple times the labor displacement, or at least they do in really good and breakout robotics examples. I'll give one here. In agriculture, yes, you would get rid of the farm, the driver, which might be 50k a year. But if a tractor lays seed and does the entire harvest season correctly, exactly when it's supposed to and exactly the way it's supposed to in every field, you see improvement in yield of the wheat, soybean, whatever it might be, that is multiple times the labor displacement. And actually, that becomes your business case. And that happens again and again in robotics. And so when you're building that case with an industrial, in this case, it might be a large farm. They're going to probably only give you credit for the labor displacement because it's going to take time to hone the operations to get the quality improvement. And so that requires both parties to agree to a long payback period. But they at least get their money back or break even, if you will, in the near term via labor displacement. And so that's been the structure that I've seen come into focus again and again. Look, in the next one, two, two, three years, we're going to be roughly a breakeven operation where we're going to pull labor out and put in autonomy. And then in year two, three, four, we're going to really hit our stride in the operations. And we're going to see this productivity or quality gain, which is going to give you a two, three, four X ROI. This is a five-year journey. And if you're willing to go on this five-year journey, this will likely have economics that are an order of magnitude better than anything else in your portfolio. If they don't believe that, you've really got to assess if this is kind of a right connection. But that's what I've seen kind of come together and work in robotics. And I think organizations that have been through robotics before make great next customers because they lived it. And so that'd be one thought.

Steve: 34:39

What you're describing resonates very close to home. So I live in the St. Louis area, but I live off Route 66 in Illinois of a my next door neighbor is a cornfield and right i am seeing the big tractors i see them all the time i've not yet seen an autonomous one out here in the fields but what i know for a fact is that the humans needed to get the fields ready to continue to to farm them they are they are not as readily available it's a talent issue around here right now my my son is 20 he's got friends that That he went to elementary school with, that high school, that at 11 and 12 years old, they're out with mom and dad working, driving the tractors, et cetera, et cetera. And I think it just resonates when you talk about agriculture and sort of I look up and I look out at the fields and it's really, really real life, at least here for where I'm at. Corporations, short-term investment, long-term investment, given the speed of all of these tech, How do I know what the best long-term use cases are? How do I know the investments that I'm making today are going to be the right ones tomorrow, given a lot of uncertainty, given the speed of innovation, et cetera? How do you navigate that?

Sam: 36:15

Yeah, this is something I get asked a lot by the corporates who are typically thinking on a longer time horizon than our robots. My view is this. Everyone here is familiar with scenario planning. We very often will scenario plan regulatory scenarios into the future and see how our strategy holds up under those. I think it's important for corporations, especially those who have lots of physical touchpoints where robotics could play a role, to have future scenarios of what robotics can and can't do and to test their strategic decisions, whether that's market entry into a new segment, like they're going to serve a new use case, or it's entry into a new market, like maybe they're going to enter Asia to reduce costs, like essentially offshoring something. I think those big corporate decisions need to start to be analyzed through the lens of what can and can't robotics do, much like we put them through the lens of what will and will not the regulatory environment look like in the future. Because The variety of outcomes is that large in robotics. You might hold a corporate perspective that humanoids will fall below cost of some of the more specific robotic use cases. That's going to drive a certain set of decisions. You might have a belief that, for example, some of the on-road stacks are actually going to permeate through some of the off-road use cases like construction and ag, etc. And that's going to impact the decision you might make to enter those markets. So I think you have to have almost like a scenario of record or two that you test each of your market entries against. That's the way that I've kind of recommended it and seen some companies do it for both robotics and other breakthrough technology. So it's not like this is totally new to the concept of robotics.

Philipp: 38:11

So in the conversations I have with corporates and robotics companies, there's of course always the question, you know, do we need to own this at some point, right? Do we actually need to own the technology? Is it okay to partner? Should we maybe do it ourselves still? Is there still enough time to use our own engineering expertise? How should corporate think about the classic built by partner decision M&A in this space?

Sam: 38:38

Yeah. So when I think about a corporate that would make an M&A decision, I quickly eliminate corporates who are going to use the robots in their own operations. I've often seen, like, you've got some examples, right? Like Amazon acquiring Kiva years ago. That was one of the few instances where a corporation that was going to implement those robots in their own operations had sufficient scale to warrant the buy decision or even the build decision if they'd gone that route. I think that's few and far between. And so what you're left with is the corporations who might sell forward the robot, right? So think like an OEM who might buy an autonomy company to integrate that with their vehicle and then sell that forward to their customers. Those are kind of the use cases where I think about or have discussions around robotics M&A. And in that space, I have been quite hesitant of build. I think the engineering profile that you need to do robotics tends to differ pretty dramatically from what's in-house. I think only the largest players who've been willing to play the long game have seen success here. So think like Caterpillar and John Deere that these folks have built internally and scaled it and successfully deployed. But outside of the five, ten mega names, I really think down to partner or buy. And I think buy is a very interesting one because of the market dynamics. So you have a lot of robotic companies that get funded. And these are big bets that the VCs are making, and they have a typical VC success and failure rate. And what that creates is a lot of, let's call it, robotic program starter kits. These are a robotic company that has managed to bring together 10, 20, 30, 40, 50 really bright minds who span the robotic stack, which is an immense one. You see software engineers, firmware engineers, hardware engineers actually You literally need one, two, three, four people of each of those kinds to make a robotics program come together. And so it presents an opportunity to essentially acquire that team, bring them into your corporate and begin robotics development for yourself. And I think that that has been the most successful or viable path for M&A in robotics is large corporates buying these programs in a box and then running them towards their desired goal of robotics development. robotics that lines up with their vehicles, etc. So that's the one that typically excites me the most. We talked earlier about partner. I think there is a tremendous amount of room here. This is actually what most of the companies I work with. It is structure and partnerships with these OEMs. But we kind of covered that earlier. Where that doesn't work, I think this acquisition model that I described is kind of the next most common.

Philipp: 41:34

Thank you for that. But if you think about your work and you bring robotics into these companies, and workers do see that robots are taking over their job. They may do the job better, maybe faster, maybe better quality. What is your advice to CEOs? How do they prepare organizations and the workforce for this machine-robot interface, the human in the loop? How do they have to think about that to really make it a meaningful workforce implication Yeah,

Sam: 42:09

I think this is a great question. is I'm going to lose my job. These robots are just going to take it over. I might as well figure out what I'm going to do next. And I think the reality is when you look at how robots run in these organizations, an incredible number of jobs emerge and they require new skills. So anytime you deploy robots, every three, four, five robots, you need a remote operator, someone who's watching them and interjecting when they become confused or are about to do something wrong. That's a new skill set. These robots have much more complicated maintenance schedules than basic vehicles. That's a skill set that needs to be trained up. And when you look at the number of people who have to do those roles, and then you go look at the attrition rate of that employee base at the company, it actually becomes less about people losing their jobs and more about how do we balance the transition. You're going to naturally attrit more people than you can bring robots in and train people up to manage them and that math needs to be explained to the workforce so that they understand no it's not the case that you are now not desired it is actually the opposite you're needed more than ever and we're needing to ask you to do more and that typically comes with more pay and here's our plan on how to train you up are you willing to go on this journey it's essentially a leveling up journey and it's on the ceo it's on hr leadership it's on the operations management to trickle that message down to the teams and then execute the strategy against it early, early on before the first robot shows up so that people actually feel like it's real and we're doing it. That's been what I've seen succeed.

Steve: 44:24

Sam, I want to close sort of giving advice, getting your advice on what you're seeing, things gone right, things going well, etc. Before we get to that question, I'm curious, Corporate leaders are navigating both robotics and AI right now. They're very symbiotic. I think some might say they're very synonymous. Do you view those technologies as different? And do you think the adoption of those two technologies are going to take similar or different paths?

Sam: 45:02

So this is something I've been thinking about a lot in the last six to 12 months as you've been watching all the news as the generative AI movement has kind of graduated from just like chatbot to an actual agent that's running entire workflows. And I think what you're seeing is the process by which something like an AI agent is going to be integrated into a business's operations actually has a lot of parallels to the problems that robotics has been working on for integration in the last five to 10 years. The things we talked about earlier in the show. So if you take an AI agent, you have to go into the business and find the version of the workflow that it's well equipped to take? What are the stack of mortgage papers that this agent can actually solve? And what are the stack that are too complex for it? Let's peel those off. You then have to update the process, right? There are a way that humans do it today that is not going to work for the agent. And that has to be thought through with the working team. And then you have to actually go find measurement measures for this that are different than what you were doing for software before because you're doing effectively labor displacement here and productivity and quality gains. And so you have to come up with a whole new set of KPIs that are pretty foreign to SaaS software, but pretty well known to robotics. And so as I looked at, it really made me think that the playbook that robotics companies have been running for the last five, 10 years to get this technology integrated into the business is going to probably look quite similar to what these agencies And I think there's a lot for the teams that are integrating agents to learn from the successes and the hardships of robotics companies that have been integrating what is essentially a physical agent, right? So you're just kind of, you're essentially replacing a workflow with either physical or digital robotics. And I think there's 80% similarity in what good looks like as far as the steps you need to undergo. So I'm super excited to see these two worlds kind of merge and hopefully for AI agents to achieve a little bit of an acceleration because there's a bit of a playbook or groundwork that exists here.

Steve: 47:14

Yeah, absolutely. Ultimately, when I hear you share your point of view, I think fundamentally it gets to where we started the conversation in terms of value delivered. And I think the value delivered from a physical robot showing up at your door, delivering groceries is a little bit more tangible to your point. And the AI and the software behind it is a little less tangible in terms of the value and And it's an important, critical role into making that experience happen.

Sam: 47:42

Yes. Yes, agreed.

Steve: 47:44

So why don't we close with some advice, maybe your point of view. You look around, you clearly are spending a lot of time and expert in this area advising companies. What do you think are investors getting right? And what do you think are missed opportunities? What do they need to adjust? I don't know if they're getting it wrong, but what are areas for improvement? Let's put it that way.

Sam: 48:07

Yeah. So I'd say on the right side, a really exciting trend that's happened over the last two, three years is you've watched a lot of investors move their VC dollars into more tractable robotic use cases, right? For the last 10 years, a lot of the money was going into on-road autonomy, which we made incredible strides. A lot of the use cases that are more tractable today are more tractable because of the work there. But you had a bit of a starvation for funding for the use cases that could be solved in two, three, four years versus the use cases that took five, 10, 15 years. And in the last couple of years, you've seen a shift to some of those nearer term use cases, things like agriculture, construction, warehouse automation versus on-road driving and on-road trucking. And I think that that's healthy for the ecosystem because you need a good mix of near term, admittedly smaller TAMs, but still very large. and long-term huge TAM solutions. And I think right now we're in kind of a good mix. And so that's exciting. I'd say on the areas for improvement, I think there's a chance for VCs and CVCs, I think, can kind of carry the torch here, by the way, to go a layer deeper on the technical due diligence. I think robotics is a space where if the technical promise comes to fruition, the economics many times follow. And so the question is more about what are the technical bets I'm making? What's the parlay I need to hit here? What are the three breakthroughs that need to happen for this technology to be able to run this robot? And I think if VC spent a little more time there, either with partners who are deep in the technical space, or if you're a CVC, you have the ability to tap into your corporate network, right? This is one of the reasons why I think following CVCs in robotics investment, is an excellent opportunity is because the CDCs have access to a bunch of engineers, a bunch of technical talent, a bunch of operational experts. And those parties can actually help complete really rigorous technical due diligence that'll result in a better pool of investments for the CDC and VCs who partner with that. Absolutely.

Steve: 50:30

Sam, thank you for joining us on the show today. Such an interesting topic. I think certainly our audience are going to be very interested in your point of view. But this navigating of what I think is just terrific technology and at the speed at which it's coming and how you balance that with delivering on shareholder value, it is such a complex, but I think an exciting time for AI and robotics and from corporates all over the place in this space.

Sam: 50:58

Yeah, no, thank you all so much for having me. I completely agree. It's a big lift to get this technology working, but it ends up being well worth it. And so it's a really fun journey to take with companies.

Philipp: 51:09

Well, thank you so much for your time, Sam, and great to have you on the UPaaS team. And yeah, looking forward to doing some incredible work over the next couple of months and years. And thanks for being here. Absolutely. Likewise. Thank you for having me.

Steve: 51:29

Thank you. To learn more, visit upath.com. That's the letter U hyphen path dot com. And to catch up on all of our episodes, visit upath.com forward slash podcast. I'm Steve Smith. As always, thanks for listening to Inside CVC. We'll see you next time.