Amanda Razani: Hello, I’m Amanda Razani with Techstrong.ai. Excited to be here today with Artem Kroupenev, and he is the vice president of strategy at Augury, where he oversees Augury’s AI-based machine health performance and digital transformation solutions. How are you doing today?
Artem Kroupenev: Great, thanks. It’s a pleasure speaking to you today. Thanks for having me.
Amanda Razani: Glad to have you on the show. Can you talk a bit about Augury and the services that you provide?
Artem Kroupenev: Sure. So Augury is a billion plus valued company, one of the first companies to achieve that kind of unicorn status or valuation within the industrial AI space. We provide machine health and process health solutions that are powered by AI. And our approach has been for the last 10, 11 years to really make sure that functions of reliability and maintenance as well as now with process health, the functions of process engineering and operations get the most out of AI technologies and are able to do their work dramatically better. So that’s what Augury provides. We have a full stack solution around machine health, predictive maintenance from sensors all the way through to extremely accurate algorithms and teams that make sure the customers go through the transformation of utilizing AI enabled technologies to make sure that their machines don’t fail and their processes run better. And we’re doing the same thing now around process engineering to make sure that processes across different manufacturing segments are automatically centerlined and can produce sustainably can produce better in many different aspects.
And we work with some of the largest manufacturers in the world, mostly process manufacturers. For example, eight out of the 10 largest beverage companies work with Augury, six of the largest CPG companies, three of the largest food companies, two of the largest pharmaceutical companies. And list goes on and on across different segments, whether it’s building materials or chemicals, oil and gas, and a number of other industries. This is where we provide a lot of value to make sure that companies are able to produce better, produce sustainably, produce without failure.
Amanda Razani: Wow, wonderful. That’s a great segue into our topic of discussion today, which is: Manufacturers integrating AI. So how are some of the world’s manufacturers currently utilizing AI technology for their daily operations?
Artem Kroupenev: We actually conducted a report recently using a third party where we interviewed about 500 executives in manufacturing. So this is quite timely as a question. And when we think about the key use cases that have emerged over the last few years. Augury has been helping lead some of these use cases. But those use cases within the manufacturing space are divided into three categories. The first one is reduction of unplanned downtime, optimization of asset care, empowering maintenance and reliability leadership to utilize data and insight in terms of their decisions as opposed to utilizing the traditional time-based methods. So that’s the first area of use cases around predictive maintenance and machine health. The second one is around process health, what we call process health is optimization of quality yield and throughput around production processes. And this is a growing area and there are some dramatic improvements with utilizing AI and optimizing those processes and process engineering in general.
And the third one is optimization of the cost of materials and the cost of essentially waste and energy. So the area that goes deeper into sustainability and managing the resources in order to be able to produce more effectively, more sustainably. Those are the three overarching use cases that are very apparent and we’re helping lead some of these use cases to scale with manufacturers, right? So it’s not enough to just say, “Here’s a promising use case that where we can potentially apply AI. We’re starting to see some results.” All three of these, and especially the machine health use case is already at scale across hundreds and even thousands of facilities globally. So this is… The scale matters when we talk about AI in manufacturing.
Amanda Razani: Can you provide any specific use case examples with specific companies and what were some of the benefits that were seen after implementation?
Artem Kroupenev: Absolutely. I mean, we have a number of public publicly available use cases on our websites with specific customers. Some of them, one example is PepsiCo, which is one of, I think, today the largest beverage company in the world. They were able to avoid 10 million pounds of wasted food. So essentially reduce waste as well as increase the capacity of production, but avoiding downtime. And that is a tremendous benefit across the supply chain. So it’s not just the immediate bottom line savings for the company. It’s also top line because some of the companies, CPG companies, food companies actually sell everything that they produce.
So without essentially building new production lines, new production facilities, a company like PepsiCo can increase their capacity by some percentage points, which is tremendous. That means avoiding building a new production facility on all the cost as well as the environmental impact that entails and their knockoff effects throughout the supply chain for a company like that where we’re talking about reduction of fuel for trucks around transportation, better management across their supply chain in terms of their capacity, better decision making in terms of what materials to… or what materials to use where and what product to produce and what facilities based on how efficient and reliable they are.
And of course then overall efficiencies in terms of their spend on maintenance, on spare parts, on the way they utilize their workforce. And also on how effectively can they upskill their workforce because people who are no longer are essentially putting out fires within the production process and running after different issues that they have that they were previously blind to. Now they have the foresight to understand exactly which piece of equipment, how are they going to fail, when are they going to fail, how to fix it, and therefore they can utilize that time to learn new skills, to create better maintenance practices, to create better strategies. And that has been tremendously beneficial to them. So PepsiCo is one example. We have public use cases with companies like Colgate-Palmolive, we have DuPont is another example and many others.
Amanda Razani: Wonderful. So it sounds like many key topics and issues that are positively impacted by this technology.
Artem Kroupenev: Absolutely, yeah. Once you have… What’s interesting about the AI, like any revolution in the technology… We can talk about the nuances. We’ve kind of used this blanket term AI, but the implementation of this type of technology into a use case that’s critical to manufacturing, like maintenance and reliability, unlocks an order of magnitude of better capabilities. Now, it takes time for these capabilities to trickle down and kind of disseminate throughout the organization. But let’s say a 10 or a hundred x impact on your ability to predict machine failure and avoid that failure could be up… could translate almost immediately in up to 20 or 30% in terms of your overall efficiency. And as those use cases become wider in the scale, more and more than the overall efficiency of manufacturing grows, not just by dozens of percent that can grow by hundreds of percent in the future.
Amanda Razani: So that brings me to another question. From your point of view, what are some of the biggest roadblocks or challenges that manufacturers have when they’re trying to implement this new AI technology?
Artem Kroupenev: I think there are three areas. We talk about mindset, we’re talk about skillsets, and also we talk about the legacy of technology. Maybe start with the third one. When we think about software in the world of digital goods and intellectual property, it’s much faster to change things and to introduce new technologies, new capabilities, and also to adopt them because the legacy infrastructure is not as ingrained and it’s not pure metal as it is in the manufacturing space. So you have to take that into account, the legacy platforms, the legacy systems that exist within manufacturing and also the practices that we’ve had over the years to maintain and optimize those systems are a lot of the times a barrier that you have to work with in order to implement a new technology utilizing AI. So in some cases there’s a lack of infrastructure, for instance, the right data infrastructure in order to be able to implement technology, which we solve for by introducing our own sensors, our own how hardware to make sure that the data is right.
But in other cases, it’s actually the way those teams have been operating for a long time and the legacy processes that they’ve had in place and the way people have been trained that need to be changed in order to adopt this new technology. So in some cases it’s actually faster to implement AI or something like machine health utilizing AI in an industry where there hasn’t been any legacy practices around reliability or they’re very minimal over time. And then industry like let’s say beverage production or wood products and some other industries versus an industry like oil refining where they’ve had 40 years of legacy practices around making sure that reliability is on par. And so in some cases it’s faster to go into an industry which is not as traditionally mature as others. So that’s in terms of technology.
The other piece is, I kind of mentioned mindset a little bit through that, but the mindset and skillset piece, I often hear that people talk about we need to upskill the workforce, we need to make sure that people are digitally savvy. We have people who’re they’re in their 40s and their 50s and they’re not as digitally savvy and therefore we have to train them. I think that’s part of the answer, but I don’t think that’s the whole picture. When you think about the consumer world, there is no manual for Instagram or TikTok or any other application that you can just go and start using. You don’t have to be trained to be able to use those technology because they are user-friendly and they’re built to be intuitive. And so the honest and really the responsibility for making sure that people within the manufacturing space are upskilled in order to utilize these technologies, at least half of it, if not more falls upon vendors like Augury to make sure that the software that we’re providing, the capabilities we’re providing and the service that we’re providing is as user-friendly as humanly possible.
And even push the boundaries of that to make sure that option is there. Nobody wants to come in out of college with an engineering degree and going to work at a plant and start using software that’s 40 or 50 years old. It’s incredibly demotivating to do that. And so that’s the approach that we take is that we are responsible for upskilling the workforce through our products, through our services equally as much as the people who are operating the manufacturing floor. So I think that’s going to be the skillset.
And then the mindset is that really people manufacturing typically just in general as an industry this is an engineering industry. This is an industry that’s based on science and technology from the onset. Now the mindset is generally if you already get that to a point where it’s operating efficiently, it’s efficiently, don’t change it. Let’s not break it. Let’s design something to operate for 30, 40 years and not change it rapidly. So I think that mindset with the introduction of digital, especially AI, not just software, but AI technologies has to change. It has to change quite rapidly. We need to get back into a mindset of continuous improvement that’s not periodic, but actually iterative and is constant.
And we need people in leadership to embrace that and understand that. And so we need to allocate a certain percentage of the time for people in manufacturing towards experimentation, not just have that as a couple of roles within the organization that are scouting for technologies and trying to implement, but everyone’s time needs to be actually start to be allocated towards experimentation, towards utilization of new things. So that improvement becomes a part and parcel of how we work in manufacturing. Once we’re starting… and we see that across a number of our customers, a number of partners we worked with, they’re kind of embracing that mindset and they’re seeing tremendous results. And this is also based on our survey that we conducted, difference between leaders who are embracing AI technologies and manufacturing and they’re kind of doing that have that mindset of experimentation and those who don’t is dramatic. It’s 30 40% higher efficiency improvements and we’re just getting started. So I think that’s just as a message around mindset that’s important. And this is probably the most difficult thing to overcome.
Amanda Razani: Great points. And it brings another question which you touched on a little bit, and that is how does Augury technology stand out from other AI solutions?
Artem Kroupenev: Well, today we don’t talk about technology as much as we talk about the results of what we provide. I think that we have been both fortunate to work with great customers, great partners, but also to have been fortunate to hire some really smart people over the last 10 years and grow the company by an approach that the technology that we have is cutting edge. And we utilize every branch of AI that’s applicable to understanding and predicting machine failure, as well as process efficiencies and some other aspects. But the technology, or the approach to technology or the technical stack, is not as important as have we been able to scale it with customers? Are they seeing dramatic results and improvements? And can we then take that understanding knowledge of scaling and apply it to other customers and other industries. And we can. And so today, when we talk about what differentiates us is that we’re a leading company within the machine health space.
We’re kind of helped create that category of AI driven machine health and predictive maintenance. We have a great number of customers and we’re growing a very fast clip and our customers are seeing hundreds of millions of dollars of value realized for them. They’re becoming better, they move faster, becoming more sustainable. And the use cases that Augury’s helping them drive around AI are actually spearheading other areas of digital transformation around their organizations. But that has to be driven by an approach to value as well as we are setting the tone and the metrics of what does it mean to have an AI driven machine health capability. And some of those metrics include you have to be over 99% accurate in your prediction because people are actually relying on the accuracy of those algorithms in order to make decisions day to day. And some of those decisions, if the algorithms are not accurate, can be actually very costly for the organization or the environment.
So we take that responsibility upon ourselves. And we do the same both for machine and process health in terms of those capabilities. The other thing that we have, I think that differentiates us is the approach we took from the very start in investing, not in technology or a layer of technology, but investing in the function or the people, the users that we serve. So when we come to solving the machine health use case, we look at the reliability engineer and we look at the maintenance person and we understand what is the hardest problem that they’re facing day to day. And maybe the most difficult problem to solve is, well, when are my machines going to fail? Are they going fail? How do I make sure to prevent their failure? And the whole job depends on it. And we built the software, the hardware and AI capabilities to solve that problem, to make that function better.
So the full stack solution, not partial solution, but the full stack to help those people do their job better is what really differentiates our approach. It is a harder path to take as a technology company because you have to build so many different elements. But once you are successful in doing that provides tremendous value. And it’s a very quick to be implemented, very quick to scale and it just works. So in that case, the results speak for themselves. And we have a similar approach to process engineering and operations within continuous production processes where that’s exactly what we’re doing, solving the hardest problems for process engineering through that full stack approach, looking at the user, what they need to do in the jobs to be done. And then longer term, and I think we’re already started to build a lot of the infrastructure for that is then combining those use cases together.
So it’s not enough to just solve the machine health and maintenance use case or the process engineer use case. Manufacturing is a system where all the components are interrelated and correlated. And that is also the power of AI is you can solve for specific functions, but then understanding how they play together, how do the machines affect the quality of your product? How does the process and the throughput affect the way you do reliability? How does all that affect the way you utilize energy and how does that affect your supply chain planning? Those questions are what manufacturing executives grapple with every day. And the promise of AI is that we can actually tie between those insights in a much better manner than you have previously. And so that is where we’re working towards. We call that kind of tying between the different insights and the way they influence each other we call that area, that category production health, and we’re building towards that with a good number of our customers already.
Amanda Razani: Fantastic. So last question. What trends and advancements do you foresee in the integration of AI within the manufacturing sector in the next year?
Artem Kroupenev: I think until now, one of the biggest advancements is of course around conversational AI, generative AI technology. And that does change a lot of things, or at least has a lot of prospect to change our perspective, at the very least on how we implement these technologies. And here’s what I mean by that. The technologies around optimizing processes or around optimizing the health of equipment are what we call narrow focused AI technologies that utilize signals to solve a specific problem. They can have physics based into it, baked into it domain expertise. They can create a whole stack around that specific problem, but they’re narrowly focused and that’s why they can be brought to a point where they’re actually really good and highly accurate at solving these problems. But that’s just half of the picture. When we think about an enterprise or a manufacturing company, machine signals, ERP signals, signals around your materials, your energy use, the way things move within the physical sphere of manufacturing.
That’s just half of the picture. The other half of the picture is interactions between people, the wisdom that we have, the knowledge sharing, the work orders that we create, the conversations that we have with our customers with the market. So all that conversational information gets lost when we focus just on the physical components. And the promise of gen AI is that we can actually combine a narrow focus into these domains of manufacturing and the signals that they provide, the insights we can get there with a more general look at how the manufacturing organization operates and what the relationship it has with the market, with its customers, with its suppliers and so forth.
And that is not just a kind of an engineering function that we’re thinking about. We’re thinking about more of an executive function where AI can serve as a co-pilot or maybe co-creator in actually running the business of manufacturing, not just part of the process of manufacturing. And that has an impact on the interface and the way we think about integrating. We call that a hybrid intelligence approach. We integrate human expertise with AI expertise and knowledge, and then we have human judgment integrated with the ability to plan in a very good way utilizing AI. So I think that is really what the future holds, not just for manufacturing for many other industries, but for manufacturing. I think that change will be quite dramatic within the next few years.
I think that is one thing. The other piece is that obviously most manufacturers have their eye on AI and on scaling some of these use cases first, integrating them and scaling them. And I think that trend is going to accelerate and continue and we’re fortunate to be part of that trend and even elite parts of the use cases that we talked about.
Amanda Razani: Fantastic. Well, thank you for coming on our show today and sharing your insights about AI in this sector.
Artem Kroupenev: Thanks so much for having me.