Speaker 0 00:00:00 I'm max Reynolds, I'm the co-founder and CEO of Symbio where we're enabling the next generation of industrial automation.
Speaker 1 00:00:12 Hello everyone. And welcome to the robot industry podcast. My name is Jim Beretta and I'm your host. And as you heard from max, we are happy to have max Reynolds on from Symbio robotics. He is based in San Francisco, California, and I, uh, would like max to go ahead and introduce himself, max, who are you and how did you get started?
Speaker 0 00:00:33 Um, so, um, the co-founders somebody robotics I, uh, got started, um, actually, while I was in the PhD program at UC Berkeley, I've been working in robotics, um, for just over a decade now and saw an opportunity to bridge the gap between, you know, some of the technologies that are being developed in academia and real world applications and industry.
Speaker 1 00:00:56 Thanks, max. Um, I'm going to, I'm here. I'm going to ask you some questions and, um, uh, hopefully have a bit of fun today. Um, so how has Symbio making robotics easier to use and is it just any robot? Is it collaborative robots? Is it industrial? Which ones are, were you using?
Speaker 0 00:01:12 Yes, we're very focused on industrial robotics and industrial applications, primarily in manufacturing and assembly. And there's really two components to what we focus on the first is advanced capabilities. And then the second is ease of use. So we've started by working with some of the largest manufacturers in the industry, some of the largest users of industrial robotics, and really at that sort of level in the market, what our customers are most interested in is advanced capabilities. You know, what can I do with my robot today that I couldn't do yesterday? And for us ease of use is an important component of making that sort of technology more broadly accessible either within a customer or an account or within a market segment
Speaker 1 00:01:58 Doing most of this with software, not as much hardware, correct.
Speaker 0 00:02:02 Exactly. Yeah. We're an industrial robotics company, but we're really focused on building a software product. This allows us to be compatible with many of the top industrial robotics providers, some of the brand names that have been present in the industry for some situations decades, and really help support our customers in terms of leveraging that automation and more efficient way
Speaker 1 00:02:25 And doing some of the, um, research ahead of our podcast. Uh, I see that you are doing a lot of integration using force torque. And so what else is in your tool?
Speaker 0 00:02:35 So in terms of capabilities, we really focus on three core capabilities. The first is forced feedback, which is the equivalent of a sense of touch for a robot. The second is computer vision, the equivalent of a sense of sight. And the third is machine learning. We leveraged these capabilities in order to enable new types of salute automation solutions that weren't previously possible.
Speaker 1 00:02:57 So you can come in to a existing automation system and add vision where we need to add vision at force torque, where we need to add four stork. Correct.
Speaker 0 00:03:07 Exactly. Yeah. This is a software layer, you know, that you can think of as application infrastructure or middleware that enables industrial automation applications to leverage advanced capabilities, you know, along the lines of force feedback, computer, vision, machine learning, generally AI.
Speaker 1 00:03:23 And so let's talk about AI and machine learning. And because of course it's a, it's a very popular topic and I think there's a lot of confusion in the industry about it. What's so important about deep learning from your perspective.
Speaker 0 00:03:35 I think deep learning is, um, you know, one important component of AI, which leverages sort of deep neural networks, um, fairly complex machine learning models in order to do function approximation. So basically in order to using data model physical phenomenon, that would be otherwise very difficult to model deep learning. Isn't always necessary in terms of supporting and industrial automation deployment. Um, deep learning represents a lot of complexity. Um, there are, you know, much simpler machine learning models that also tend to work pretty well depending on the type of physical phenomenon phenomenon that you're trying to model.
Speaker 1 00:04:14 And so do you have to pair like AI and deep learning, or can you separate those two?
Speaker 0 00:04:21 You know, what I would say is deep learning is a subset of AI and AI is a pretty broad umbrella term that covers a pretty broad variety of disciplines, whether that's computer vision, machine learning, and, you know, each of those categories can be if sub-segments had into the specific categories of models that, that you choose as well. Um, so where deep learning really comes in and has, you know, made a really big impact is actually in the computer vision community where, you know, computer vision is a very data intensive problem and deep learning has demonstrated some pretty exciting results, especially in academia, in terms of modeling certain phenomenon, you know, visual phenomenon and images that you couldn't do your a sickly or with a more simplistic model.
Speaker 1 00:05:08 And how critical is it to your customers to be also adopting this in real time? Is that kind of the essence of what you do?
Speaker 0 00:05:15 Yeah. Real time, you know, is kind of the name of the game in robotics, right. Um, and when you think about controls engineering, you know, the goal of, you know, building a controls algorithm is to control a robot in real time. And so, you know, where, where we have really focused on innovating is in terms of embedding some of these machine learning models, whether that's a deep learning algorithm or something more simplistic, um, like a simple linear regression algorithm into these control systems in real time. And that's where we've been able to add a lot of value in terms of either improving the efficiency of an automation solution that was underperforming or enabling a new automation, automation sort of automating something that wasn't previously automateable.
Speaker 1 00:06:01 Do you ever have a target when you come into an application, let's say you're doing the first one, which is your, you want to increase efficiency of an automation system and can you get like two or 3% or what, what is kind of your end game when you come into improve an existing automation?
Speaker 0 00:06:17 Yeah, I mean, it, it, it often depends on the complexity of the solution, but we've been in situations where we've generated, you know, uh, over 25 X improvements in terms of the convergence rate or the, the rate at which the automation solution converges to ideal cycle time. We've also observed situations where just through a retrofit we've shaved, you know, somewhere between 10 and 15% off of cycle time, just with, you know, uh, an upgrade to, to the software layer that's associated with the system. So we're seeing pretty drastic improvements, particularly in some of the more sensitive assembly processes that we've, we've engaged. And there's, there's a big push in the industry right now in final assembly, which is often referred to as the final frontier for automation. And, um, there are a lot of, there are a lot of solutions that are going into this final assembly area that by definition tend to be a little bit more finicky, um, just because it's more cutting edge, um, more cutting edge automation solution. And, uh, we've, we've, we've demonstrated by retrofitting some of those types of, um, work cells with our software. We've been able to demonstrate performance improvements along those lines.
Speaker 1 00:07:33 So I'm assuming when you talk about assembly automation, that you're really talking a lot of work in automotive, are you doing any work in any, any other sectors? Are they on your target?
Speaker 0 00:07:43 Yeah, my, my background is actually my first job was in aerospace manufacturing. Um, I worked for Northrop Grumman on the F 18 production line, and that was actually the first place where I was exposed to industrial automation. So yeah, automotive is the, the golden standard, if you will, for industrial automation, the sense that industrial automation technologies are being leveraged in a very broad way, you know, in some situations across global manufacturing footprints with over a hundred facilities. And so we have done a lot of business in automotive, but we also have engaged with customers in other verticals, like aerospace, heavy industrials, um, et cetera.
Speaker 1 00:08:26 And so how do people engage with you? Do they call you up and just say, Hey, listen, we've got this underperforming arm or you, you get, or is it a bit of word of mouth or
Speaker 0 00:08:33 How does that all work? Yeah, we do have a website, um, SIM view S Y M b.io. And, um, we do get inbound requests through our website. We often connect with our clients through LinkedIn, but really the, the primary way that we like to engage with a new client is through a facility tour where, you know, our, our emphasis is on connecting production stakeholders that are supporting automation equipment, you know, day in and day out and really focusing on supporting them.
Speaker 1 00:09:02 So what kind of, uh, I, I'm assuming you probably have a couple of clients or a couple of industries that you can talk about as use cases.
Speaker 0 00:09:10 Yep. Yeah, absolutely. I mean, I think, you know, when you, when you kind of take a step back and look at where automation lives today, the killer use case for industrial robotics technology has kind of been welding right. And welding in the automotive industry. There's a lot of pick and place, and certainly a lot of automation and paint, but when you walk downstream in the manufacturing process where we see a lot of opportunities coming into Symbio is final assembly, which is less than 5% automated compared to 98 or 99% automation or welding, um, particularly in automotive. And so, you know, we've really focused on supporting assembly operations, primarily in final assembly, both in situations where you have stopped stations and you want to improve the performance of that stop station, or in situations where you have a moving line assembly tasks that needs to track a vehicle or a sub sub component moving down the production line, using computer vision, then leveraging other capabilities in order to actually perform an insertion task.
Speaker 1 00:10:14 And max, do you sometimes work with automation integrators or do you work with end customers or how does that work?
Speaker 0 00:10:19 Yeah, we we've primarily engaged directly with end customers, but we've partnered with integrators is especially as we've deployed these solutions into production settings. And so, you know, we're, we're based in California, we're primarily focused on software development or 42 person company today. And, um, and where, where, you know, it's really a win-win for us to partner with an integrator is when, you know, there's a direct relationship with the plant. There's a team that has experience with the PLC programming template or the controls template that's being used in the facility we can partner up so that Symbio focuses on the software development component in our business model is really focused on licensing. And then we can hand off some of those integration activities or the vast majority of, of those integration activities in a mature relationship to, to an integrator.
Speaker 1 00:11:08 So in doing this work, you obviously create a quotation. You come in, you look at an opportunity, you say, Hey, listen, I think we have some gains here in this, in this cell or on this line. And then you do a quote and it's that quote, kind of a fee for service, or is it a monthly service, or how does that work for you for you being more primarily in software?
Speaker 0 00:11:29 Yeah, so we are focused on software licensing and we typically provide two options to our customers. The first is a software license associated with a five-year term paid up front. You know, typically if we're engaging with a more centralized organization, there's a cap X budget. And that option is usually more, um, interesting to those folks. And then the other option that we provide is recurring license across that five-year term paid annually, which is what we leverage when we're engaging more directly with the plant that has a budget more structured around an a P and L more structure out op ex.
Speaker 1 00:12:06 And do you, are you dealing with, uh, CEO's or project managers or plant managers or this, or the chief technology officer, like where do you touch or do you touch all of them
Speaker 0 00:12:16 Kind of, kind of all of the above, um, you know, what we've found is, yeah, especially when we're talking about new technology and automation that, you know, our customers that are, are adopting that technology more, most successfully tend to leverage sort of internal partnerships between a centralized organization that's structured around either production engineering or advanced manufacturing, and then, uh, you know, working with stakeholders from those groups in partnership with plants or aligned management. So the folks, um, who are actually in the F in the factories day in day out and supporting the operations, and, um, we found, uh, you know, internal partnerships between these groups tend to breed the most success in terms of adoption of new technology.
Speaker 1 00:13:05 And so where do you see the future for Symbio? Do you just see more of the same, or do you see, uh, smaller companies, bigger companies, where do you see yourselves going?
Speaker 0 00:13:15 Yeah, we've, we've really started out working with some of the largest customers in the segment. And I think if, for me, it's a natural starting place because you know, these, these clients are the most mature in terms of adopting new technologies and rolling them out more broadly. You know, I do see technology over the technology over time becoming more accessible, um, through some of these initiatives in terms of ease of use, and in terms of identifying use cases that are much more broadly applicable to other segments and other, um, uh, sort of customer, uh, sizes. But, you know, I think we'll also see a situation where even, you know, at the top end of the market with the largest customers, you know, we're going to see it drastic improvements in terms of the efficiency of automation and existing segments. And then also, you know, for example, in assembly, uh, much higher levels of automation and where I get really excited about, you know, in terms of thinking about the future is we think about how these organizations are going to engage in automation and how the model for controls engineering robot programming is going to evolve over time.
Speaker 0 00:14:25 You can talk about ease of use and accessibility in the context of smaller customers, being able to leverage the advanced capabilities that are being leveraged at the top end of the market, but you can also have a conversation about how a broader set of stakeholders within those organizations, um, can advocate or engage with, um, automation in a broader, uh, at a broader scale. So I think we'll see, um, developments in directions.
Speaker 1 00:14:52 What are the things you mentioned earlier in the podcast is, uh, is, is kind of the algorithm and how important is it for some of these arts or companies to actually invest in the data algorithm that they're almost creating themselves?
Speaker 0 00:15:08 Yeah, I mean, it's, it's an interesting question with AI in general. Um, and you know, this kind of circles back to machine learning, deep learning where, you know, how much value is there in the algorithm compared to the rest of the process and what we've found in terms of deploying automation solutions, particularly in settings like manufacturing, where robustness and performance are extremely important is refining that the engagement from the developer and the controls engineer, the production stakeholder in that process is often just as important, if not more important than the algorithm itself at the end of the day, somebody has to develop that algorithm, right? And that developer has some intuition about the process that they're encoding in that development. Um, but in order to be able to build that intuition, having data visualization and cleaning that support that intuition development upfront, it has a really drastic effect in terms of the output of the algorithm once it's deployed in a production setting. So I would say the algorithm is important, but the process that the developer engages in, in, in terms of developing and deploying that over them can often add just as much if not more value.
Speaker 1 00:16:28 That's very interesting. Thank you for that. Um, so your perfect customer is assembly automation, probably complex mix of parts, uh, going into that final mile of automation and they're, they're less than perfect automation or they're building a new, uh, is that your kind of your bulls-eye client?
Speaker 0 00:16:46 Yeah. You know, we've, we've traditionally, we've engaged with clients who are already leveraging, um, automation at some scale, and we're often involved in helping them identify new ways to leverage automation or improve the, the efficiency of that automation. You know, I think, you know, as we, as we see some of these trends in collaborative robots, you know, we're going to see more SMBs leveraging automation and building up that sort of expertise. But yeah, today we've, we've really focused on some of the larger accounts within the segment.
Speaker 1 00:17:20 Uh, thanks for all this max. Uh, how can people get ahold of you?
Speaker 0 00:17:24 So we do have a website Symbio S Y M b.io. And, uh, we're also Symbio robotics on LinkedIn. So feel free to connect with us or, uh, or send us a message.
Speaker 1 00:17:34 Great. I would like to thank and acknowledge our partner. The association for advancing automation, 83 is the umbrella organization for the RIA, the AIA MCMA and eighty-three Mexico. These four associations combined represent almost 1300 automation
Speaker 2 00:17:50 Manufacturers, component suppliers, systems integrators, and users, research groups, and consulting firms throughout the world that are moving automation forward. I'd also like to thank and recognize our partner painted robot painted robot builds and integrates digital solutions. There are web development firms that offer SEO and digital social marketing and can set up and connect CRM and other ERP tools to unify marketing sales and operations. And there are painted robot.com. And if you'd like to get in touch with us at the robot industry podcast, our email is the robot Institute
[email protected]. And you can find me Jim Beretta on LinkedIn. We'll see you next time. Thanks for listening. Be safe out there. Today's podcast was produced by customer attraction, industrial marketing. I'd like to thank my nephew, Chris gray for the music, Chris Coleman for audio production, my partner, Janet, and our sponsors a three and painted robot.
