Gal Shaul, Co-Founder & CTO at Augury - Versatile Machine Health AI

(Joel Beasley at 00:00:03) Hello, my friends. Today we're talking to Gal, the co-founder and CTO at Augury. And we discuss how Augury's machine health AI is able to prevent 80% of unexpected downtime, how COVID has accelerated years of innovation down to a matter of months, and how to make space for different ways of thinking to get the best ideas possible out of your team. All of this right here, right now on the Modern CTO Podcast.

(Gal (Guy) at 00:00:34) Here we go.

(Joel Beasley at 00:00:36) This is the Modern CTO Podcast.

(Joel Beasley at 00:00:47) Well, let's get into it. Take me back. How did you first get into technology?

(Gal (Guy) at 00:00:53) Take you back. Well, it's going to take a while, but let's maybe start. I think I've been interested in how things work since I remember myself. It was a VCR back in the day with my dad, trying to figure out what cables go into where and then trying to figure out other things. And a lot of things at my parents' place that are already broken because I tried to understand how they work and could never assemble them back. I was always into sports as well, and even there I found some techy sports. I was in fencing for many years, and there was a lot of technology that's going into that.

(Joel Beasley at 00:01:33) What kind of tech is in fencing?

(Gal (Guy) at 00:01:36) Who touched who? How hard did you touch your opponent? What got there first? It's a lot of microseconds type of things that only computers can detect that a referee could never see. What exactly happened? So it relies a lot on technology. If you ever watch fencing, it's amazing to play the sport. It's a bit hard to watch because it's so fast. You watch the rerun. You usually don't understand what exactly happened there. But that was a part of my life then.

(Gal (Guy) at 00:02:06) I was a naval officer for a few years in the Israeli military, and then I thought of what I want to do next or what I want to do when I grow up. And growing up has a lot of different meanings. But for me it was trying to understand how to build things that matter and trying to take a lot of things that I've done before and try to connect them together. And started working more formally in tech after a few years at the Technion in Israel, which is a software engineering school in my case.

(Joel Beasley at 00:02:40) Oh, cool. So what was your first major job in tech?

(Gal (Guy) at 00:02:44) I worked in a digital cameras company. We were manufacturing microprocessors for digital cameras. Everything that does JPEG at first ended up being a lot of signal processing for how we produce cameras. So a bit before iPhones came out and cameras on phones became really good, digital cameras were a big thing. And I worked on algorithms to identify smiles, to identify waves, the kind of before-selfie type of pictures where you put the camera, you go out, someone waves, and it takes the picture.

(Joel Beasley at 00:03:22) Oh, nice.

(Gal (Guy) at 00:03:22) So a lot of signal processing and image processing. I started there as an intern, a bit like you shared earlier. I started as an intern. I said, "Hey, I like signal processing. It's a lot of software. It's very close to the hardware and how things work." So I get to understand how to build things and started there. It was a lot of fun. The company was great as well.

(Joel Beasley at 00:03:44) So you were working on the image processing algorithms on the software side or the actual processor chips on the hardware, or both?

(Gal (Guy) at 00:03:53) It was a bit of both. It was a bit of the same. So it was a microprocessor or a dedicated processor that was designed for things that cameras do specifically. So you had to compress the JPEG. You had to do a lot of signal processing as well as software development. So I was on the software development side. There was an algorithms group. Today you don't do that anymore, right? But there was an algorithms group. You were given a spec to the software of how it should work or how the simulation works, and then you implement that to the best of the capabilities of the hardware and software that you have in place depending on the type of camera that you're running it. It was a lot of fun and a lot of hard work and bits and bytes, literally.

(Joel Beasley at 00:04:35) That's cool. Yeah, I had never really thought about how much processing a digital camera actually does because I think of it as just like a camera. Until this past year, I've done a lot of video work with some mirrorless digital cameras, and they overheat a lot. Wow. And it's like an issue. And I hadn't even thought about that as a problem until it started happening to me. And I was like, "Wow, these are just little computers with a lens. Of course they can overheat."

(Gal (Guy) at 00:05:10) Yeah, it overheats. And in the past, I don't know, 10 years ago, maybe 15 years ago, a camera would overheat because it was trying to have a sports mode, right? To take 30 pictures in a row and then do all the compute and then store them. It had a lot of heavy compute it had to do, and if you didn't calculate it properly, it would overheat. You almost can't touch it if it didn't work. So the hardware and software connectivity in cameras had to work from very early on to be functioning at the best of the capabilities because there's so much condensed in that piece of hardware. Of course, today it's just one of those things a camera does on your phone. But it's a lot of the way we even look at phones today is kind of a camera with a communication device more than the phone capabilities, right? So it kind of shifted to new industries and transformed in really cool ways.

(Joel Beasley at 00:06:08) Yeah, I swear that's the thing they hype up the most on every new phone release is the camera. It's not about the phone.

(Gal (Guy) at 00:06:17) It's not about the phone. It was actually a really cool example of looking at the innovator's dilemma, where if you were from the camera industry, you would look at cameras and phones like, "Oh, they don't have three CCDs. They don't have this. They don't have that. That's kind of a different technology." And every six months, or the pace of generations of phones, they became every three days, right? Depends on how fast you're going there. They became a bit better, a bit better, a bit better, way better than regular cameras, not seeing anyone else in that race, right? But you could have seen them kind of starting from the bottom, being a very basic camera to being the standard of what cameras are today. Even expensive or very high-end cameras today, they consume some technologies from that trend that happened with mobile phones and then implement it back to the cameras ecosystem.

(Joel Beasley at 00:07:11) Yeah. Well, all right. So we're not here to talk about cameras for an hour, as much fun as that would be. But so right now you're working on a company called Augury, right?

(Gal (Guy) at 00:07:23) Yeah.

(Joel Beasley at 00:07:24) So how did you—you're the founder. How did you start that?

(Gal (Guy) at 00:07:30) Yeah. So I was working in this medical device company and signal processing there, a lot of hardware and software connectivity there as well. And Saar and I have known each other for ages at that time, right? We'd been talking about startups for five years before we started Augury. And kind of every class we took at the university, we were talking about startups, went to meetups, compared books, compared notes, even decided on our first job based on what is the right path to start a company. We didn't get to the same conclusions, but always had that in mind then. Somewhere between this medical device company I was working on, where you start seeing problems arise at software but it's actually not a software problem, it's a hardware problem. And something in a pump doesn't work, but you don't know that. So you end up triggering a heat alert or something kind of very late in the process. And Saar was working at the time at Intel in the chip design area of the microprocessors in Intel. But he has a background in physics and electrical engineering and was very interested in voice and sound in general and machine learning, and felt there was a lot of missing opportunities between everything that the world learned about image processing and speech recognition. But sound as itself wasn't as explored at the time.

(Gal (Guy) at 00:08:58) And he was looking at a lot of applications there. Somewhere, something clicked. He came out with the notion that we're ready and we can start something. We started speaking about that and figuring it out. And then a few months later, we had already quit everything else and started working on Augury, which became machine health for a lot of different applications that we never dreamed about.

(Joel Beasley at 00:09:24) That's really cool. So can you give me the overview of what Augury focuses on today?

(Gal (Guy) at 00:09:29) Yeah, sure. So for the last 10 years, we are building a company that helps people rely on the machines that matter. Now what machines matter to them and how they're going to do that may vary, but we provide superior insights into the health and performance of machines that help people make products, support life in general, provide some services around us, right? So we work with industrial enterprises that are in the manufacturing space and help them improve the health of their machines that they use currently in specific markets to avoid unplanned downtime and improve their maintenance. Based on the sound of the machine, we can tell what it's doing, how it's doing, when it's going to break with a few months of alerts in most cases. And then people repair them and get less surprises. It eliminates almost 80% of the surprises in a manufacturing line today.

(Joel Beasley at 00:10:30) That's really interesting. So you use the sound that the machine makes to evaluate the health of the machine at any given time?

(Gal (Guy) at 00:10:39) Yeah. So you can think about it if you drive your car, maybe not an electrical vehicle, but one with a motor and a drive. If you have a squeaking belt, you would know it's a squeaking belt. It doesn't matter if it's a Toyota or a Mercedes. A squeaking belt sounds like a squeaking belt. So we use those patterns to identify the main malfunctions that may occur. We look at it in a continuous manner. So we have sensors that connect to the cloud and listen to the sound, the vibration. Today, a bit more signals that the machine produces. And then look at the trends of what's happening there, and therefore we can tell what's wrong, how much time it has, and when it's going to be a catastrophic failure.

(Joel Beasley at 00:11:24) Okay. So when someone comes to you, you kind of go into their factory and put some sensors on their machines, and you can kind of, I guess in a sense, turn any device into an IoT device that's capable of reporting its current health and whether or not it's in need of maintenance. Is that accurate?

(Gal (Guy) at 00:11:47) Yeah, that's accurate. So you can look at a production line that is comprised of hundreds of different units, right? If you're trying to make beer, right? So you have beer, it arrives somehow. Then you have bottles. Then you have the cork. Then you have the, I don't know, the plate that it has. Then it goes into a box, right? So it has a few components, but it's hundreds of machines in the process, and each and every one of them can fail. Some of them, let's say 40% of the machines are actually critical. If one of them is not working, the entire line is not functioning, right? And it can be in manufacturing beers, it can be in manufacturing medicine, it can be when you try to make toilet paper or diapers or toothpaste, right? It's kind of all of these are reliant on a lot of things that work simultaneously in order to be manufactured in the fashion that we're trying to make products and get them to shelves. And then what Augury does is, yes, we put sensors in each and every one of those devices and try to alert and give the team that's working on keeping their health insights to what's going to go wrong, how are they going to fix it, and what's the urgency. Do you need to stop everything now and repair it, or can you wait for next Monday when you have two hours of shutdown and you can easily replace different things that need replacement?

(Joel Beasley at 00:13:07) So how do you go from all the raw data from the sensors to insights on the health of the machines? I imagine with so many different machines, like within manufacturing, machines get very niche and different from each other from plant to plant when they're creating, when they're making different things. So do you have some AI running that is able to learn the sounds of the different machines and vibrations that are necessary to convert that data into insights?

(Gal (Guy) at 00:13:41) Yeah. So Augury, in the sense of our data science and AI, is a very hybrid company. We use mathematical models and very modern AI and networks that we can dive into later if we're going to go really geeky and techy, which is phenomenal and interesting. And on the other end, we use a lot of physics. We bring people that are subject domain experts in machinery that have been around for years around those machines, coming from predictive maintenance, and have seen patterns of those machines, and we marry them together. So we try to marry the physical elements and physical understanding of what happens to a machine. What does that sound represent in friction? What does it represent in things that happen to a bearing and kind of real understanding of machinery? Tie that with AI and with networks that we're using in order to understand what is the rate that this malfunction is happening. So getting a prototype in our market is relatively easy. You connect a machine to a vibration sensor. You look at the raw data. If it's a hallmark malfunction, you will see that right away. How to get to the nuances where you have a very quiet system that alerts you when you actually need to do something, that's where our core expertise comes into place. And that's kind of reliant not only on is it really the problem, but is it the right time to repair that? Is it the right time to fix it and not overload you with more information than you ever want to consume?

(Joel Beasley at 00:15:12) So do you also interface with existing IoT devices? Could you plug something that's already got connectivity into your interface and take data from that as well?

(Gal (Guy) at 00:15:24) Yes, we can. It's not our go-to solution in order to get started, but it is something that we do later on in order to gather more insights into the production and manufacturing of the entire operation. For the one-machine type of thing, for a bearing that went wrong, we use our own data or we use very specific data that we want to have. If we want to get to insights that are more holistic on the way production is happening and how to improve it at different levels, we're doing that by integrating to different sensors and different things that are already measured, already are in the cloud, and we try to match that. In general, in manufacturing, a lot of the problems in digital transformation initiatives is that it takes a lot of time and requires a lot of contextualization of the data. You get a lot of sensing, but what does that mean? How does that tie into the production? We provide very fast ROI on the program because it doesn't integrate at first with any other system. We have dedicated sensors that we put in, and then we connect them to the cloud. Then we bring them to you and not try to mash all the IT and OT, operating technologies of the world, into one place, which is very important but doesn't bring you ROI after a minute. And a lot of our customers have seen a lot of integration projects that have been nice in theory but very hard to prove that it actually changes behaviors and changes their bottom line. So we're trying to get a holistic approach to the product and then very fast to go to changing the process that the company is running. These are traditional companies, so it can be challenging, but we are getting very good at that.

(Joel Beasley at 00:17:11) That's cool. Yeah. I didn't think about how big of an advantage that is of using your own sensors and having your system just ready to go with that. And also, as you just mentioned, the more traditional companies, I imagine having your sensor setup that you're able to bring in and immediately evaluate the health of any machines, that gives you an advantage working with those traditional companies as well that may have a lot more legacy machines in their manufacturing.

(Gau (Guy) at 00:17:45) Yeah. It's really hard to go only on new machines in our market because there are millions and millions and millions of machines out there that are pretty reliable. They're pretty reliable in the sense that they mostly work. Mostly work. When you have a car and it's reliable, it's great. You just keep driving it. When you have hundreds of machines per line and then you have hundreds of lines in your company, it becomes a statistics game. How many of them are going to fail today? How many of them are going to fail tomorrow? It becomes kind of big numbers, and you can avoid 80% of the surprises. I can't believe that this just happened. Well, tell you what, you have 10,000 machines. It happens to your company every day. Let's just eliminate the level of surprise of the people on the ground and help them understand that as they go through their day to day, which is pretty challenging as well.

(Joel Beasley at 00:18:38) So one thing I always like to ask founders of the company, where did the name come from?

(Gau (Guy) at 00:18:44) Great question. So Augury is kind of a prophecy in Latin. It exists in English, but less used. Augurs of the Roman Empire were looking at birds, and based on their behavior, were trying to tell you when is the right time to go to war, when is the right time to crop your field or harvest your fields. And they were looking at things that were not really related to each other and try to provide predictions from those signs. So that's a bit of what we're trying to do at Augury as well.

(Joel Beasley at 00:19:22) That's really cool. I actually had to do a double check, double take when I saw this on my calendar today, because a little while ago, we had a company called Auriga, which spells their name pretty similarly. And actually I do want to ask you though, based on that interview, they're a custom software development firm for outsourcing development and doing team augmentation and software. And their CEO was telling me about how outsourcing used to be somewhat taboo, but throughout COVID that stigma has kind of lifted and more people have been open to outsourcing, and that's probably going to stay as an industry trend overall. And I'm curious, what are some changes that have been brought to your industry by COVID? And do you think those are here to stay?

(Gau (Guy) at 00:20:21) Yeah. So our industry has definitely gone through digital transformation from being, hey, we just added the chief digital officer and we're trying to figure it out and there are many pilots and we're trying things out in digital transformation in manufacturing, to, hey, this is real. This is happening. This is here to stay. We're doing it. And it happened for a variety of reasons, but a lot of them was it's nice in theory, but the transition was hard. The people on the ground, maintenance managers and sometimes plant managers, depends on who they are and what company they work for, were a bit reluctant and a bit skeptical on the actual effectiveness and credibility of new solutions. They've been around. They know their machines. They know their site. They have good instincts of what's going on there. And they're like, yeah, we'll try it out. Maybe some of us will. But it didn't become this trend. The industry suffered from real pilot purgatory around a lot of IoT and digital transformation initiatives. During COVID, three things happened. Right? The workforce suddenly is not on the plant anymore. Not all the team is there. So if a maintenance manager now is remote, they need a communication device with the team that are there. They can't just walk to the line and give orders and go back. They need something that is within context of this specific line or specific machine to communicate on, and they don't have their own ears and eyes to see that. So it started with, hey, we need more tools. We need to manufacture more. Our industry had to manufacture two, three times more in most of our customers during the beginning of COVID, with a third of the team on-site because of social distancing. And it created the demand for more technology. And then the workforce is changing dramatically. 17% of manufacturing workforce is retiring in those two years, right, between 2020 and 2021. And it was actually accelerated because the workforce that was a bit on the older side of the map were retiring early, or it was more risky for them to come back with COVID, so they got retirement plans to be effective earlier in their career. So it got to, alright, now we have a new workforce. They don't have those instincts. They're more keen to use new technologies, digital technologies. And, actually, they need it in order to make quota of what they're trying to manufacture. All of that is there at stake. Right? None of it is going back. In terms of outsourcing, not outsourcing, it's not the right industry. Right? Manufacturing is still very traditional. It is in a way that we're looking at the machines that they have. And some of their vendors, which we work with, want to provide it as a service. Right? Don't worry about buying a pump. Buy pumping as a service. Don't worry about buying a chiller. You need to buy, actually, cooling. That's what you're trying to achieve. So let me give it to you with the right sensing, with everything attached to it, and we'll provide the service that is attached to those machines. It's kind of a leasing for industrial of everything that you need. Who owns the equipment, who is in charge of repairing that, that can also change over time.

(Joel Beasley at 00:23:55) That's interesting seeing that everything as a service model just penetrating every industry. We're never going to own anything by the end of the century.

(Gau (Guy) at 00:24:05) That's probably true.

(Joel Beasley at 00:24:08) So, earlier you mentioned that your time to ROI is really fast. How do you actually measure the ROI?

(Gau (Guy) at 00:24:16) Yeah. So a lot of it is done by our customers. They know what an hour of downtime is really worth for most manufacturing facilities. So if you make X amount of product in an hour and that product wasn't produced, the direct cost is that you're going to have less things that you can ship to the market. Now with commodities, it means that someone will buy something else. Let's assume that you were talking about toilet paper. Right? So you need to fill 10 shelves with your toilet paper. And I'm in another company, and I have to fill another 10 shelves. Right? Let's assume that you weren't able to provide those 10 shelves, and they only have my products. If someone comes to that pharmacy, they'll just buy whatever is there. No one's going to wait for another toilet paper. Right? It's kind of, yeah, I prefer mine, but I'm not going to go on a hunt to look for it. Probably same for diapers. It's almost the same for a beer. Right? Yes, they didn't have my beer. I'm upset for a minute. Then I'll take another six pack and I'll go and whatever to my friends that I promised I'll bring beer to. And I think that's a lot of it. They miss on the go to market and they miss their targets because they're not manufacturing fast or efficient enough. Other than that, there are a lot of things that get broken and you may tighten a bolt today or replace an entire motor if you wait for two weeks and it's going to run out. Right? And so they calculate the ROI on lost production and equipment that was saved earlier. Now I'm saying it as if it's simple. It takes a lot of calculation and time and complexity to get to those ROI calculations. But most of the time, they know the formula before we came there, and we just add, hey, were you able to identify that earlier? What would have happened if you didn't know it? Right? And they need internally their internal marketing or their internal campaigns to make sure that these programs are successful includes those success stories as well. So we help promote that internally to help them create that change from within, which we believe a lot of digital transformation is actually about people.

(Joel Beasley at 00:26:40) Yeah. That's cool. I mean, whoever made the purchase from you is motivated to justify that they made a good purchase to their company. And so they make those calculations, and then you get to take advantage of that as, hey, look how good it did.

(Gau (Guy) at 00:26:56) Yeah. So the person who made the purchase, usually, they're worried about adoption. It's like, yeah, we tried these new technologies. The plants didn't want it, or the people didn't understand what it is good for. Right? So we have to show them first and then enable them to promote that the adoption is really easy and it's really high among the people that they work with. And they don't have to enforce it, quote unquote, from the corporate, but it actually is a part of what their people want to do, what the teams want to do. Right? It's kind of a bottom up, if you want to call it that, from the plants and people want to use it more. And then it's a good sign for the corporate that it's actually moving forward because a lot of the software that was bought in the last few years, it's just sitting there. And we are trying not to be a part of that trend.

(Joel Beasley at 00:27:49) Yeah. Absolutely. So do you send people to do on-site training for your software after a purchase is made?

(Gau (Guy) at 00:27:57) Yeah. So we're a high touch model in general at Augury, pre COVID. Today, we're high touch, but most of it is done remotely. So we have customer success that is on Zoom 12 hours a day with customers trying to help them figure it out. We have training that happens remotely. We created a new training department that was planned to arrive later, but we are training people to do installations themselves of those sensors and of the IoT devices. And we train people to use the platform, and we use a lot of digital technologies to help push the success on the facilities as well as on corporate. When needed, we send people on-site usually for the installations and to get the service up and running. Most of our customers interact with the platform, or we have at least one or two people on the plant level that are using that every day, multiple hours a day to try to figure out what's happening with their machines and how to improve them.

(Joel Beasley at 00:29:03) That makes sense. I should have worded that differently than, do you send people out? Because, obviously, there's a lot that can be done just over Zoom.

(Gau (Guy) at 00:29:13) No. I think in this industry, I think the question was right on because in this industry, it wasn't acceptable. They expect the people to come and close deals on site and to come to train them on-site and to come if there is a problem to speak to them personally. And we had people traveling everywhere. They closed the gates, and no one that's not from the plant is allowed in. And then they had to adapt, and we were already there. We were actually helping them transition more to move to digital types of communication, asynchronous communication, and trying to make sure that a decision can be made from you and then transmitted to your team later. So a lot of it is something that we help them go through as a part of leading them through the change. The communication and feeling comfortable in a setting like we're in now was a bit of a challenge at first. Today, they feel it's kind of, that's the main thing that's happening. So it became a reality and everyone adapted.

(Joel Beasley at 00:30:13) Yeah. I mean, I'm sure now that they're used to it, from what can be done over Zoom, probably just as good or almost just as good as face to face. And they're probably saving money because you're saving a ton of money on travel expenses, and you don't have to price it as high. So it's good for everyone in my opinion.

(Gau (Guy) at 00:30:33) Yeah. It's good for everyone. I think there are a few things that still need to happen in person. I'm seeing that internally for within our company as well as with our customers and for them as well. Some things, it's very hard to quantify, but some things do require people sitting together trying to brainstorm and figure out a hard problem in creating the relationship and creating the transparency level that's needed and the confidence level that's sometimes needed. So it takes a minute, especially when we're talking about physical products and facilities that now need to shut down in order to replace a bearing because you said so. It's a big decision for a lot of people. But most of it, yeah, it happens in a different way or remotely, and it's actually working. In a lot of cases, it works better. Right? Communication's better. It's documented better. It is.

(Joel Beasley at 00:31:28) So, a little while ago on the podcast, we had on this guy named Zane Bond from a company called Keeper Security. And they're a password management tool. They do more than that, but that's their consumer facing model. And this guy was super smart, and you can tell that he just spends a lot of time thinking about security. And so when I saw your company coming on, I just kind of had security on my mind from that episode. My first thought was, all right, this kind of seems like IoT. Every IoT device is a possible threat for someone to breach and get into and control these machines. But hearing about how you actually go about monitoring the machines by listening to them rather than having devices that can actually control them, I feel like you've reduced the threat of each endpoint a lot just by taking that approach. But I guess aside from that, how do you view cyber threats at the endpoints of your machines with having them all connected in some way?

(Gau (Guy) at 00:32:45) Yeah. So cybersecurity or cyber threats, as you said, are a big topic in our field, and it starts with education. Is a device that doesn't control your machine actually a threat or a system that helps you check that no one is playing with your data on the operational data that does control the machine? Because if your controller is messed up, then you have another platform that is impartial, and you can see that it actually does what you told the controller to do, or you would have discrepancy and then trigger an alert on that IoT. I think we're in a market that even the move to the cloud took time and still takes time. Right? We have conversations with IT and information systems in each and every one of our customers, and that's a big topic. Do they already have a cloud strategy? Are they going to have their own cloud application store, or are they going to use public cloud? Right? Do they approve to work with Google, Microsoft, or Amazon? Or is that just one cloud that they're approving and all the apps need to go in there? So there were a lot of questions and a lot of fear in manufacturing in general and kind of moving to the cloud. What is it going to do? How many threats are there? What's riskier in kind of going to the cloud or keeping operation as it used to be? And I think that the transitioning that happened through COVID is really fast. Right? Things that we thought that are going to take two, three years, just six months later, a lot of our customers already had everything in place and an understanding of how to do that. And it's a part of the way we assess a segment or a market to go into. Are they ready to go to the cloud? Because it means that they have a lot of things in place, including some kind of a strategy of how are they going to operate or what do they want to achieve and not just an aspiration. Right? So having a strategy and not just, hey, we really want to go there is really helpful for us to go in.

(Joel Beasley at 00:34:51) Yeah. So you've mentioned a couple of times about how we're kind of coming out of COVID, people have really accelerated their transformations. It seems like a pretty exciting time to be where you're at, kind of implementing transformations. What is something that you're really excited for for the future of your company?

(Gal (Guy) at 00:35:11) Yeah. I'll start with the past because you asked and it came up. So I'll share that when COVID hit, we had all of our installations canceled in a day. We had a lot of things happening. And then within a month, two things happened.

(Gal (Guy) at 00:35:24) We couldn't install, and then we had to figure out or implement a new way of installing our platform. But a lot of business, or booking in that case, came through the door. So a lot of people wanted it but didn't know how to implement it. So we had a lot of bookings and we had implementations that couldn't go on because we couldn't get anyone on site. So we changed basically all the strategy of how to deploy the platform, and now it can happen really fast.

(Gal (Guy) at 00:35:55) I think going from that anxiety state of, "Hey, these are all things that we wanted to do in the future, but now we need to do it now. It's not a two-year process. We need in three months to be able to do it completely differently than the way we were set up." Today, this is our biggest advantage. We know what to do there, and we know how to do it fast.

(Gal (Guy) at 00:36:14) And I think we're excited about being able to grow in three different dimensions at the same time, right? So it's an aggressive growth model where we grow geographically, and we're going to be installed in many new countries in the upcoming six months. We're going after new markets, and we're going after new types of machines. So even within the same market, going after more machines than what we're installed on today is really exciting for us.

(Gal (Guy) at 00:36:41) So going in those three dimensions at the same time in order to really make this transformation impactful for manufacturing is really exciting.

(Joel Beasley at 00:36:51) And that sounds like going through that change within your company to get so much quicker sounds like such a daunting task. It was probably really overall helpful to you guys to have such an insane forcing function, like a pandemic that was just like, "All right, you have to do it." And yeah, I mean, that just sounds like you did it right and are coming out stronger than before.

(Gal (Guy) at 00:37:16) I think that a lot of leaders in this area, you could tell kind of what background they're coming from. So for some of us, wartime is time to flourish. Yes, you're going to have maybe post-trauma later, but it's very clear what you need to do. Focus is at its extreme.

(Gal (Guy) at 00:37:36) We changed all the communication structure within the company to make everyone extremely focused on the one problem we have to solve. And then solve that and move to the next one and then move to the next one. And it puts you—you can't be in wartime all the time because then innovation is a bit hurt, or very hurt, depending on how much stress you add to the system or add to people's minds and hearts. But for a while, all your great ideas or everything you thought of is going to come true now because you can narrow it down to what is blocking us now from helping our customers.

(Gal (Guy) at 00:38:12) What is blocking us from helping products go out of the door and provide people with what they need today? And that forcing function was really helpful in that. So we were able to move almost a step function ahead. We were at the right time, the right place. And we had great investors that were backing us up to be able to go through that change.

(Gal (Guy) at 00:38:33) And today, we're in a totally different challenge, right? There is good product-market fit and all the cracks are ironed out for that product, or most of them, right? There are tons of problems in scaling anyways, but the basic ones are out of the door.

(Gal (Guy) at 00:38:47) And then, how to move from one product to a portfolio of products is a very hard one, especially if you go to industrials, enterprises with hardware and IoT devices, right? So how do you move to a few offerings combined or a portfolio of offerings? That's the main challenge we have through the growth now. How to keep that level of focus we had during wartime in the pandemic to how to build for scale.

(Joel Beasley at 00:39:10) Yeah, that makes a lot of sense. So I see we're coming up on the top of the hour. Do you have a hard stop at the end of the hour?

(Gal (Guy) at 00:39:16) I don't.

(Joel Beasley at 00:39:17) Okay, cool, because I just want to get into a couple leadership questions before we wrap up. How would you describe the culture at your company?

(Gal (Guy) at 00:39:26) So I would say that we invest a lot in culture at Augury, and it's a very innovating space. I think it's a company that really takes to heart the way things are being done internally and externally, right? We look into our customers. We understand their process. We understand where they are. There's a lot of empathy to where people are coming from with a lot of drive to change, right? So a lot of ownership and innovation are a part of that. There is a lot of autonomy in Augury.

(Gal (Guy) at 00:39:57) So tons of cross-functional teams across the board, and people can get a lot of context. So we share a lot. So there's a lot of transparency. And then in return, we get people's ownership over their own strategy, their own plan, and then the execution in being able to figure it out. As the offering is very wide, and it's sometimes complex to understand.

(Gal (Guy) at 00:40:23) So we need people that are experts in their domain that are able to make decisions and run autonomously, or else we'll never be able to make it happen. From what I can see, we have also a very nice team. So if you ask for help, people probably give you more help than you asked for, and after an hour, you'll just say stop.

(Joel Beasley at 00:40:41) That's awesome. So has the culture of autonomy been there since the beginning, or is that something you had to focus on and build once you were already going?

(Gal (Guy) at 00:40:52) I think it was a combination. We wanted to have autonomy from day one, and then we put it as one of our core pillars and one of our main values alongside with people first. So that empathy and ownership was there from the beginning. So we were working in squads when the company was, I don't know, 12 people. So it started in the structure, and then how fast did you implement OKRs to enable people with their own autonomy?

(Gal (Guy) at 00:41:18) I think, emotionally, it was a bit of a different story. It was hard for me to let go of some of the things and some of the teams that I, quote unquote, built. And letting go was probably the hardest lesson I had to learn. When I give autonomy, you forget that you have to actually give a lot of the things that you were in charge of and make sure that people can actually do a better job than you in being able to lead them over time.

(Joel Beasley at 00:41:42) Yeah, I mean, that makes sense. The best teams are just made up of people that are better at any given task than you are. That's why they're there. So if you could design the perfect leadership training program for the leaders you oversee at your company, what would the most important one to two concepts be?

(Gal (Guy) at 00:42:07) I think the first one would be listening. I think that for me, leaders that can lead from behind can enable teams to flourish. And for that, you need to be a good listener and actually understand what people meant and bring one more question to the table or ask a bit deeper. You usually learn way more than trying to assume and just provide an answer. I think the second one is inclusion, which is very important.

(Gal (Guy) at 00:42:34) Sometimes leaders themselves or the people they like to work with are people that have a very easy, quote unquote, tongue, right? And are very open in conversation and would start speaking right away and feel confident. And I think it doesn't mean that they're the smartest. It doesn't mean it's the right solution. It's just the one that caught the mic first. And being able to get everyone in the room to a comfortable position where they can share and think together is really important. And I think with those two is enough. Empathy and listening and the ability to get inclusion through the room. If I had to provide two pieces of training to my leadership team, that would be keep investing in those. I think the third one is being honest.

(Gal (Guy) at 00:43:20) I don't get maybe all social cues, and I prefer honesty over everything. So I prefer people to just stay very direct to each other, what they think, where they're going, how are we going to win this, and be able to have honest conversations at all times. It's really helpful as a third one.

(Joel Beasley at 00:43:39) That makes a lot of sense. It really hit home with me hearing what you said about how the loudest person isn't always the right person, or the person that's most willing to talk doesn't necessarily have the best idea. And that's why it's important to get everyone's input, because I personally am the person that is willing to talk and I'm always open to share my ideas. And I'm often wrong, like a lot of the time. It comes up because I write music with my friends, and we're working together on something.

(Joel Beasley at 00:44:17) And I just come up and just say aloud the first thing that comes to mind. And it's really important to kind of pry into some of my other, like, quieter colleagues that are just sitting there thinking it over and coming up with a much, much better idea than what I said in order to get to the best possible final product.

(Gal (Guy) at 00:44:42) Yeah. I think it really relies on, goes back to how people learn. So some of us are online and need talking in order to figure ourselves out. I don't know what I want to say, but I'll start talking and then figure it out while I talk. So it's learn by talking. So you need to make room for those people. You also have people that are writers, right? They need to go home, think about it, and write and come up with a great solution, right?

(Gal (Guy) at 00:45:06) Does it matter that it happens today or tomorrow? Well, just because we scheduled the meeting for the day, so the decision can't wait for tomorrow for you to think about it a bit longer and come back with something that you wrote? Usually, that's not the case. Some of us are readers, right? How do we learn? Some of us just read it way faster than listening in a 40-minute call, right? So some of us want to read it, right?

(Gal (Guy) at 00:45:29) There would be people that, instead of listening to the podcast, would just read the summary or read someone that wrote it, and it would be 10x faster to them, and they would learn or understand more, right? Some are listeners. So you want to be able to capture all those learnings: people that are more online and are going to do it on the spot, all the offline thinkers, and get really a place where the inclusion comes into place and allows them to think longer, allows them to come with their ideas later, even if it was in a text message.

(Gal (Guy) at 00:46:03) Right? Sometimes the best ideas come back, "Hey, did you think about whatever X?" And it's like, "Whoa, where did that come from? Let's cancel the meeting that we have and go with that," right? So you have to allow that to happen, and I think that sometimes takes time. And we got used to, especially pre-COVID, right, in a world where leaders were tested on those skills, right?

(Gal (Guy) at 00:46:26) What did you test leaders for? "Hey, did you speak first? Did you take ownership? Did you move the ship in the right direction?" Right? So a lot of that outgoing was required. But actually, when you look at innovation processes, they're not necessarily happening all the same. So being able to be modest and say, understand that it's not the only way to communicate, is really important.

(Joel Beasley at 00:46:47) Yeah, I think that one thing that you touched on really stood out to me, and it's that oftentimes, we kind of set a deadline for a decision to be made by the end of a meeting because that's what the meeting's for. But yeah, you're totally right. It doesn't matter whether or not that decision's made right now or later tonight if the decision that's made later tonight is a better path to go on.

(Gal (Guy) at 00:47:16) Yeah. I think there is something about the level of creativity of the—are we actually solving the problem? Do we have the right setup and relationship to make it work? That's one end of the spectrum. And the other end is, are we efficient and effective? We set out it's going to be an hour. Did we meet our goals at the end of the hour, right? I think what happened, at least to our team during COVID, is that efficiency and effectiveness became great. It's those creative moments that I'm afraid of, right?

(Gal (Guy) at 00:47:43) It's those water fountain moments that people from different teams were talking and something snapped, right? Those are the ones that we care about. Those are the reasons startups even exist, right? It's that chaos that creates innovation that is a bit missing in large organizations, or sometimes missing in large organizations. And you want to keep that and not only worry about effectiveness and efficiency or other metrics that are generally measured.

(Joel Beasley at 00:48:12) Yeah. So I guess before we wrap up, I just want to make sure we hit on everything we want to hit on. Is there anything you want to get out there, you want to plug, or what have you for Augury or otherwise?

(Gal (Guy) at 00:48:27) Yeah, I think when I was thinking about listeners of this podcast, I think that we sometimes, us CTOs or tech people in general, we sometimes fall in love with technology speaking to other technologies and getting humans out of the loop. And I wanted to remind all of us that all of these transformations are for people, by people, to people, and including them in it is really important. And sometimes it's, "Hey, how do we make sure that people are actually going to use it?" cannot be overstated in this environment. So keep them in mind.

(Joel Beasley at 00:49:06) Thank you so much for listening. And if you found this episode useful, please share it with a friend or colleague who you think would get value from it. And if you have topics that you'd like to hear discussed on the podcast, either add me on LinkedIn or send me an email, [email protected]. Every time I get an email or LinkedIn message, it absolutely makes my day and inspires me to keep going.