The AI Myth Everyone Believes!

Full transcript

Welcome back to the podcast, guys. Today we are joined by Bryan Thyken. Brian, welcome to the podcast. Thanks for having me. Let's kick this off. What do you do and how do you spend most of your time? My name is, or I'm the chief revenue officer for Sorba AI. We are an end -to -end AI ML platform that allows subject matter experts to develop models for the industrial automation space in no code. How do I spend my time? I spend most of my time working with my sales team, my marketing team, to evangelize what we do and work on our strategy on how we increase revenue for the company in a very complex and also crowded space, especially with the terminology of AI out there. You have a lot of clarification and distinction you have to make between yourself and others in the industry. And when you look at the industrial AI space right now, what is something that most people still get wrong to this date? What people get wrong. So I think one of the biggest things, at least with how our product and

technology works, is that you need to have it trained on years and years and years of data and from multiple different data sources, other customers, etc. You know, we've kind of proven out that you, one, the best data to train on is your own data that you have within your own plant or facility. And two, you don't need three years of it. In fact, three years of data is probably going to be training it on bad information. So you're really not getting the best value for the information you're providing it. So I think that's one of the big things, this kind of idea of needing years of data, tons of data. And I think the other piece that is important is also the need that the data science community. Now, data scientists are immensely important. We have data scientists who use our software all the time. But what we really focus on is the SME part.

And the SMEs are the most important part of building anything with AI. You need to have that expertise on how a piece of equipment works, you know, what the outcome you're trying to achieve, what the optimal goals are of increasing or optimizing your production and finding faults and maintaining uptime. I think a lot of people have this feeling that it's like a real data science piece. And it's actually, in our opinion, and how we've also focused on building our tool set is really around the is the value. They know how that's supposed to work. The person who's been in the plant 25 years and knows when it wiggles a certain way that something's going on. And that's the piece that we really focus on is capturing that subject matter expertise and being able to expand it exponentially across the plant. So, you know, one person can't do everything. So how do you use software to replicate that person to the best of your ability to optimize their skill set? Thank you. Thank you. This is a link in the description. You're looking for the information we have. is that right

Correct, yes. You know, work past any potential traumas you might have experienced while in the service, leveraging just the natural therapies that come from being in the outdoors. And so the work that Ben and Josh and team have done at the Veterans Outdoor Advocacy Group is immeasurable. And, you know, working with Congress to get laws passed to further that cause and get appropriate funding down to state and local organizations to help serve the veterans community. got it yeah i wanted to ask you if you see some benefits of actually being involved with that group like what's the connection between the advocacy work and the corporate strategy That's a great question. I mean, that's a great question. I think one is, you know, when you look at any sort of advocacy work you do, most of time the it's being done for free. You know, whether that's volunteering, you know, at your local school in the PTA to, you know, large national organizations, you know, what you should be bringing is your best skill set to the team. And so I look at when you get involved in outside you.

organizations that it's important to, you know, do what you do well to enhance that. And hopefully it's something that you're super passionate about. And so I think that that's really the part to, it kind of gives, it augments your own life, you know, in terms of the daily grind that you have to, you know, commercialize a product and increase a business and grow. You know, having that outlet on the side where you are making an impact in a different way, I think is super important for any professional to have, you know, like I said, volunteerism is not a, you know, join an organization. It could be as simple as, you know, just going out and, you know, helping with a community pickup or, you know, of the neighborhood or something. So I think it's important for all people to have some level of engagement into their community or into a group of people that you have you.

association with. And I think the military community in general is one where you walk into a room and the moment you may not have been in the same branch of service, but you recognize each other. And there's a shared story, even if you weren't in the same places or, you know, one is a combat vet and one wasn't. The ability to kind of immediately pick up like you've been friends forever is a really interesting and unique experience to have just in general. Going back to Sorba and the use of AI in Sorba, how are you guys currently using it? Is there any specific use case that you could share? Yeah. So really what we are is a tool set that enables the SME. So whether it's creating anomaly detection or classification models to regression models and all the way to like advanced process control where you're trying to optimize stuff in closed loop or digital twins and forecasting, we don't really drive the use case. What we are really trying to do is make

a tool as simple as possible to allow a controls engineer or a process engineer to build stuff. So we do have a ton of use cases from, you know, optimizing different refrigeration systems or, you know, optimizing specific processes than maybe a brewery or an oil refinery. These are typically driven by their end user because that subject matter expertise resides inside the facility. So we have, it really doesn't matter what the vertical is. It could be oil and energy to food and beverage to mining. We've really seen a little bit of it all. And I think what's the most important or impressive part is when you're able to take away some of the complexity of having to write Python code or understand whether or not you need a deep learning neural net or a, you know, random forest model. When you take that away and you allow an expert to dive into a problem, it's really impressive, you know, what the results can be by leveraging some of these new techniques and technologies that exist out there. So I think outside of, you know, some of the use cases, which we've seen, you know, incredible savings or, or even just efficiency pickups, I, I look at what our tools solved and just some of the features that kind of pop out from when I was a controls engineer, you know, that would have saved me a ton of time in terms of finding a root cause of a problem within a plant that I was working in. So, yeah, there's a ton of use cases, but there's

no vertical that we really focus on. And I think that that's an important aspect of, you know, creating tools that enable those people. You could hire a million people and still not cover all the expertise you need. It's too broad and every plant's different. And how you make a cookie at one plant is different than how you make a cookie here and how those recipes go together. You know, having a tool that enables that person to do the work, I think is an exceptional contribution to the industrial manufacturing and automation space. Get into. From a technology selection and adoption or just, you know, what is kind of a bigger problem that they're facing, you know, in, in the, in the facilities. bigger problem... Thank you. Gotcha. Yeah. So I think there's, there's a big problem with just expertise in, in a plant in general. One plants as they should run very lean. Um, so they're trying to, you know, reduce their operational capital, that the human expenditure to, um, manufacture things as lean as possible.

Uh, any good business should, should be striving to be as lean as they possibly can. But with that, you also run this high risk of attrition, you know, just through people becoming older. And this is an old stat, but I mean, you know, 10 years ago, the average age of an automation engineer was something like 55 in the industry. So you, you have a very senior group of experts who are already very thin, uh, you know, um, highly utilized in, in the operations. So if you lose that person, you. you also lose so much of that expertise with them. And so having the manpower, the training, the kind of the, the passing down of the knowledge of, you know, when this machine does blank, this is the problem. And you need to fix it rather than very junior people. The other problem is you just not having a lot of people going into the trades and or in general, you know, um, Thank you.

people graduating college these days. And, and this is not a new trend. This has been going on for, you know, 10, 15 years. People don't want to go work in a dirty, loud, noisy factory. They want to go, uh, work in a, you know, build code and live this Silicon Valley, you know, software startup life, uh, that it seems so, you know, uh, enamoring, but you're, we're also losing a lot of that expertise and how to make things. Um, and so I think from a, from a macro perspective, you know, technology adoption is not there for, uh, you know, trying to replace people in a, a, you know, I have a person there. It's really honestly trying to augment the people they can't find. Um, and so then when you kind of dial it down into when people started looking at technology, I think a lot of the challenges they have is there is, um, a skills gap and even identifying good technology and how to solve problems. You have this big gap of, you know, I just didn't understand what AI is, whether you're doing a large language learning model to what we're doing, which is machine learning and an AI within time series data.

You know, there's just so much noise out there. It's hard for people to separate the signal. And, um, I was speaking to a client of ours and, you know, early on in engagement, he said, you know, like I get, you know, three to five emails a week. Of some AI technology that's going to change my world. And, you know, I, can't I evaluate them all. And I, and I, I don't even know how to differentiate them. Um, and so, you know, you look at from this macro view of we're struggling to find people or retain people or people are aging out and retiring, uh, as they so well deserve after so many years of working there to getting new people in selecting tech becomes just another very complex cog. And that wheel. And so, uh, I think the proof of value proposition is, is a little broken and how they evaluate. There's just a lot of complexity in that. So it becomes difficult to pick a tool and pick to the solution that really drives value at the end. You, you are missing a lot of those key decision who would help drive that decision forward.

Got it. And you already touched up on it a little bit, but if you could snap a finger and change something about the way that building brand awareness works, what would you, you, you, you, you change? you, About you. you, you, you, you, you, brand awareness Yeah. I mean, that's. This is, that's a, that's a long podcast in and of itself. I mean, creating brand awareness is a, um, a very challenging piece and, and telling your story. Um, again, nobody buys features and benefits. And, you know, one of the topics I'm very, you know, when I talk about sales is sales and purchasing in general. If you're a buyer, it's an irrational decision. Thank you. don't buy things because, you know, it's the best product. You may think you are, but you are buying things because of kind of inherent biases. And, um, you know, you grew up and your dad drove a Ford truck. You were never going to drive a Chevy truck. It doesn't matter if the Chevy truck is cheaper, better, faster. It's not important. You have an affinity to that Ford truck.

That is no different. Even at looking at new technology and trying to brand yourself and you fall into, you know, you may have the best piece of tech. And I've, and I've asked this directly to customers sometimes, you know, we are an independent technology company. If my software was half as good, but I had the, you know, Sorba or Siemens logo on the top of it, would you buy it? And the answer is like, yeah, we, we trust the brand, even though it's an irrational decision because maybe my tech is better. So that is the hard part as a new technology company and being in a very noisy space is try to separate your signal and value proposition from all the noise that exists, as well as competing against the irrational buying patterns of, of the end user, which is going to be completely biased to, you know, I've used Siemens my whole life. You know, if it's not a Siemens product, I'm not going to use it or a Schneider or Emerson or whomever. So you, you're, I think that's the biggest part as a new technology company is how do you differentiate yourself?

on your tech alone, but on, you know, kind of a lot of the irrational buying processes that people go through. And, and ultimately what, what is the problem they're truly looking to solve in their buying process? Um, nobody ever buys, you know, I think the ROI conversation is always important, but I think it's even a little bit of a myth, you know, uh, in, in why people buy tech. You know, there's very distinct problems that sometimes don't have very distinct ROIs associated to them. And so this constant battle and give and take of, you know, what is the irrational thought that they have versus the rational one they're truly trying to solve? And how does your product or, or group or team fit into that story? Um, it's, it's, very it's challenging. It's a, it's a daily, it's a daily mental, you know, workout to try to constantly, you know, evolve. And then as you knew, new people pop up, I mean, there, there's just, it's constant trying to, to evolve that. Fantastic. Thank you very much for the time. you.