How Kawasaki Engineers Data & AI into a Competitive Advantage (Ep 012)

Data Secrets Episode 12 with Anthony Gondick from Kawasaki Engines

In this episode of The DATA SECRETS Podcast, we chat with Tony Gondick, technology leader at Kawasaki Engines, about how data centralization and agentic AI are reshaping the future of manufacturing.

Tony Gondick explains how his small (but mighty) IT team moved beyond “good enough” thinking, tackled legacy silos, and enabled business users to become empowered data storytellers. We cover:

• The challenges and rewards of consolidating data from legacy systems into a central platform
• Practical steps for cleaning, governing, and making sense of decades of operational data
• How Power BI and Salesforce Manufacturing Cloud empower non-technical users across the business
• Real-world examples of AI-driven automation for customer service and early quality warnings
• Why the hardest part of adopting new technology is shifting mindsets, not just systems

From enabling analytics on the shop floor to building early warning systems with AI, Tony Gondick shares insights any business can use to unlock the secrets in their own data.

Connect with Tony Gondick:
🔗 LinkedIn: https://www.linkedin.com/in/anthony-gondick-05719417/

Follow the DATA SECRETS Podcast:
📬 Episode recaps & insights: https://www.allegroanalytics.com/podcast
🕺🏻 Nathan Settembrini on LinkedIn: https://www.linkedin.com/in/njsettembrini/

Follow the DATA SECRETS Podcast

📬 Get episode recaps & bonus insights at allegroanalytics.com/podcast
🕺🏻 Connect with Nathan on LinkedIn
📺 Watch every episode on our YouTube Channel

Transcript

Tony Gondick [00:00:00]:
Data is a story. It's got a story to tell, and these people now are those minstrels, those storytellers that are taking this and they're telling the story of the business, of their particular team, and how they're helping the business build success.

Nathan Settembrini [00:00:21]:
Your data has secrets. Secrets that could change everything. If you only knew where to look. Welcome to the Data Secrets Podcast. Welcome to the Data Secrets Podcast, where we uncover how business leaders crack the code, find truth in their data, and turn insight into action. Today we have a special guest, technology leader Tony Gondick. Tony, welcome.

Tony Gondick [00:00:50]:
Thank you, Nathan. Appreciate the opportunity to be here.

Nathan Settembrini [00:00:53]:
Absolutely. So Tony, a lot of manufacturers talk about data and AI. Everyone's talking about it, right? But you at Kawasaki Engines, you guys have actually centralized your data, you've governed it, and then built a agentic AI on top. So I'm just curious, like for other manufacturers out there, like what do you think, or why do you think so many folks never get as far as you guys have?

Tony Gondick [00:01:18]:
Excellent question to begin with. I've spent a lot of time talking with other peers and other individuals in this industry and in this space, and, and one trend that continues to come out is within manufacturing we've built a reputation around our, our build quality, around the consistency in how our products are manufactured. And so when you ask folks to introduce some new change, a disruption to that, they get very concerned about how is that going to impact my brand and my brand's reputation, right?

Nathan Settembrini [00:02:01]:
Yeah.

Tony Gondick [00:02:01]:
And so they get into this trend where what we have is good enough. Why change if people are buying our product and, and we really don't need to involve anything else to make modifications to improve upon that. And so the conversation will generally get into then, well, if you're not changing, your competition certainly is. And so we took that mindset of our competition is always looking for ways to gain market share. And my personal belief is that what we have is never good enough. We have to continue to evolve. We have to continue to change. Because our customer and our markets are always changing. So we took the approach to continue to build on emerging technology, adopt AI as a platform, and feed it a lot of the data that we have. As a 130-year-old company, you can imagine we've got a lot of data. So we need to be able to understand the story that it's telling, and we feel that AI can help us in that. Out.

Nathan Settembrini [00:03:16]:
And so, folks never get that far because there's this resistance to change. And they kind of need someone like you who is a kind of a pioneer for change because you're never satisfied because, hey, guys, everybody else is chasing after us. We've got to keep running.

Tony Gondick [00:03:35]:
Is that right? Yeah, I would agree with that. Yes. Yes. Because as I said, the markets are always changing. What we're seeing is a lot of non-traditional impacts and changes to our markets, to our customer. And, and Amazon is the, is the best example of that. They continue to introduce different disruptions across the board. They're the ones that have set delivery timeframes. They're the ones that have set return timeframes and customer interactions. Much like Walmart and Target in developing their apps, have set precedents for what a good customer integration or interaction is. And so as you talk about different global companies and as the markets tend to blend, those interactions of the customers start to bleed into every interaction they have with every supplier, with every producer, with everyone involved in their orbit. And so as an organization, as a manufacturer, we have to keep an eye on that. And we have to make sure that we are looking at who is setting the trends in these areas and how is that going to impact us overall in a long-term perspective.

Nathan Settembrini [00:04:57]:
I love how, you know, while you wouldn't say that Amazon is your competitor—

Tony Gondick [00:05:03]:
Not directly, correct. Yeah.

Nathan Settembrini [00:05:05]:
You know, they're not making engines like you guys do. But I love that you're you're looking outside of your industry or your direct competitors for inspiration. And yeah, that's fantastic. Could you, so, you know, we talked a little bit before this about, you know, your, the business and, you know, when people think Kawasaki, I know I was thinking jet skis and motorcycles, but could you give us just a quick snapshot of the business you're a part of, the scale you're operating at and, what might make your environment uniquely hard or challenging?

Tony Gondick [00:05:42]:
Certainly. And you are right. As soon as I say Kawasaki, everybody goes to the Ninja motorcycle, the jet skis, our Ridge and Mule line of side-by-sides. But if you look at the Kawasaki global brand, it is actually made up of many different divisions across different industries. You'll find a Kawasaki division in robotics and aerospace. In precision machinery, in the power sports, which, which does the, the motorcycles, but there's also shipbuilding and even plane and helicopter building. Now, my particular division is the general purpose engines, and our primary product is a build-to-order engine for our OEMs that is a high-horsepower, small-profile engine that you'll find in, in the back of a lot of professional landscaping equipment. So if you go to Lowe's or Home Depot and you take a look at a lot of the zero-turn mowers that they'll have out for viewing, you'll see a Kawasaki engine that powers that machine. And that is what I support. We have 7,700 authorized dealers across the United States. We have a global distribution chain through Canada, South America, Australia, New Zealand, Japan, Europe. I think the, the only continent we may not touch right now is Africa. So we have all of these customers, we have all of these products that are out there, and we're running operations out of Michigan with R&D and business operations from a small office of around 150 individuals. The team that supports them, my team within IT, is only 4 individuals. And so we have to act very much in a nimble and agile method to make sure that we're, we're staying ahead of the competition, that we are bringing in emerging technologies, reducing technical debt, and contributing to the success of the business. And I'm happy to say with the choices we've made, with a lot of the effort that I've done over the last 8 years with Kawasaki, we are achieving those goals.

Nathan Settembrini [00:08:13]:
That's awesome. Yeah, I think you're in a pretty unique role because technology is very holistic in your case, right? You have everything from traditional infrastructure technology everything from Microsoft and all that kind of stuff. And then applications that the business uses every day, like Salesforce Manufacturing Cloud. And then you also have data and analytics, right? You have basically, and now AI.

Tony Gondick [00:08:49]:
Yes.

Nathan Settembrini [00:08:49]:
Right? And so a lot of times I feel like companies bifurcate that. And so what I think that that's actually a really powerful combination. I'm curious your thoughts on kind of reasons to consolidate versus reasons to separate those responsibilities.

Tony Gondick [00:09:11]:
Part of our technology stack, well, in fact, all of our technology stack was built over time as a business process became relevant and evident. Generally, you would throw a technology solution at that, and then what you would find after a number of years is you would have a mainframe over here and an AS/400 over there. You'd have a system of record and an ERP software that duplicated functionality all over the place. And after a while, the people that supported those systems were long gone. The coders that built your ERP or your systems of record are no longer there. They've retired. It's using older technology and programming languages. So how do you deal with that? And so we had to make a conscious decision knowing that technology continues to evolve and change substantially almost every 3 years. In order for the company to remain successful, how do you do that? How do you reduce the dependency on all of these individual systems? And how do you make cost-effective choices so that the company isn't just blowing all of their profit margin on IT solely? Because you have staffing costs, you've got development costs, research and development, you've got building, So you, you need to make sure that that profit and revenue is spread evenly across all areas. And so we took a really hard look at that, and after determining where we had overlap, we started to eliminate a lot of those legacy technologies and legacy applications because the newer platforms that were out there— and we'll use Salesforce as a good example of this— they've started to build natively into an out-of-box format a lot of those business operations. And we had then the added value in making those choices, the centralized data. By doing that, we were eliminating the silos and who was entering in data. We had more data consistency and we had better data accuracy because it was one set of data fields and information that was being worked on by all the teams with, with the necessary permissions instead of that system here and that system there and trying to, to tie all of those together either programmatically or through some import and export process.

Nathan Settembrini [00:12:06]:
Yeah, that's actually a great pivot into our, our first data secret. Where we're talking about kind of getting the data all in one place as kind of the starting point, right? Because a lot of companies, when they start this journey towards, hey, we grew organically, we have all these systems, and we're trying to make better decisions from it, but some systems hold some information, this other system over here holds this other information, and now it's like, I need to— I need to see a more complete cohesive picture of the business. And so they start down this path of we need, you know, business intelligence. And these days a lot of people jump to we need AI first, right? And, but I think the, you know, the lesson, the biggest lesson I've learned from you so far is around this centralization. Like this is a process we need to go through and step one is to bring that data together. So could you tell us just a little bit about that story of how you guys went through that?

Tony Gondick [00:13:08]:
Yeah, so as we took a look at AI, as I did my own internal investigation, what I was seeing was a higher potential for introducing bias and hallucination when you had multiple data sources. The less control that you had over how the data is inserted into a system and how you retrieve it back out, And we'll touch on even the recording engine that will extract that information. But the more sources you had, the higher risk that you had in some additional bias and additional hallucinations being introduced into a more global dataset. And so we had to make sure that as we were taking all of these data sources across all of these different systems, We first went and cleaned up what we could. We know data is never ever going to be 100% accurate. As long as there's someone entering in data, you're going to introduce their personal bias, you're going to introduce their human data entry errors. And so you have to, you have to really identify What is your risk tolerance? If our data is 95% accurate, are we okay with that? Well, what about 85%? What about 80%? Can we go to 60%? Those are internal business risk conversations. But once you have that, that sets the bar then for how you bring in your data. You take a look at what your tolerance is, you start cleaning your data, to that tolerance and you start then to centralize it. And by doing that, you're eliminating that silo, you're reducing and mitigating a lot of that bias and hallucination, and you have a more accurate collection of data that you can then feed into your AI solutions or really any of your technology stacks, knowing that it is fairly representative of business operations and that it has a story to tell. And that's the most important part. All data is a story. Now, how you, you build that story with the tools that you choose, with the tolerance levels that you choose, that defines whether that story is going to be a fairy tale where there's a happy ending or a horror story where you're going to see many Hollywood sequels.

Nathan Settembrini [00:15:58]:
Nice. Yeah, so I was just wondering, you know, for the audience, so I know you guys are on Salesforce Manufacturing Cloud, and so this data centralization from a platform perspective, are we talking about Data 360? Are we talking about I imagine there's an underlying data warehouse of some sort that may have not existed in the past and your team brought it to life. Is that right? So can you give us just kind of like your technology landscape just briefly?

Tony Gondick [00:16:31]:
Yeah, so we are experimenting with Data Cloud. We'll start there because that's the easier way to answer this. What we do have, because of all of the product lines and the different OEMs that we have, We decided to put all of that information natively into Salesforce and Manufacturing Cloud.

Nathan Settembrini [00:16:54]:
Oh wow, okay.

Tony Gondick [00:16:55]:
All of it resides there on platform so that we can reach in and pull out what we need to. We can do our business planning around that data. We can use Power BI as a good BI reporting tool for that information. Now, if we do have to reach out to other systems, mainly we use AWS because we can control the consumption and usage of the storage and Bedrock or any other connecting database and AI solution that might be out there. But we, we actually use a combination of MuleSoft and Data Cloud to reach out to those systems, pull that in, normalize the data, and then make it available and readable within Salesforce. But really Data Cloud and MuleSoft is a last choice only because we, we take a look at all options of how can we retrieve that data? Is it different enough from our current data structures that we need to go and we need to normalize it or rename it or manipulate it in some way. I choose less customization and more native out-of-box functionality because I think it's cheaper in the long run. And from a long-term perspective, I'm trying to reduce the overall cost of maintaining these systems. And so natively out-of-box, we feel that that's a better choice for us.

Nathan Settembrini [00:18:37]:
Interesting. Okay, so if I'm an analyst at Kawasaki Engines and I open up Power BI and I'm going to connect to the data, I'm connecting to Salesforce in this case. It's basically your system of record.

Tony Gondick [00:18:52]:
Yes. Warehouse. Okay. Yep, that's it exactly. And then if you do more in-depth reporting, you will be reaching out to our ERP. We've made that connection. We do have some products that cross divisions. And so if you have to reach out to another AS/400 or another cloud-based repository, we can do that as well and take a snapshot of that information. We always leave it in the system of record and only take a snapshot because we don't want to deal with a lot of data leakage potentially or any other risks in that data being incorrect or manipulated in a way before it gets into our reporting engines.

Nathan Settembrini [00:19:39]:
OK, great. OK, so along those lines, so data secret number 2 is letting the business do analytics. And so you guys have gone through many iterations from Salesforce reporting to Tableau. I believe you tried Tableau for a little bit and then eventually landed on Power BI because adoption mattered more than elegance, I believe. Is what you said in the past. And so, yeah, could you just tell us about that decision and how you came to that conclusion?

Tony Gondick [00:20:15]:
So in our business, in our industry, a lot of our internal business partners are mechanics. They are experts in gas-powered engines. They may not necessarily be experts in technology standards or the latest and greatest emerging technologies. And so we had to find a hybrid solution to meet all levels of technical savvy when it comes to technical skills.

Nathan Settembrini [00:20:46]:
Okay.

Tony Gondick [00:20:47]:
You're right, absolutely. We started with Salesforce reports. We basically hit a ceiling in that. We exceeded what the native software could do. Moved into Tableau because we did have some teams that had used that in the past, but we found that adoption and that learning curve was very difficult for a lot of our business operations. And so we had already moved to Microsoft 365 and we had access to Power BI. We turned it on for a couple of more technically savvy advocates in our organization, and we asked them to pilot this, try this out before we made a commitment. Into a lengthy or expensive licensing model. What we found by doing that smaller pilot with those technology advocates, they more readily adopted or adapted into that menu system. They could very quickly see where they needed to go and using the application to build the reports. Because it was primarily built on Word, Excel, PowerPoint menus. So they just translated from Word and Excel, Office 365, over to Power BI. And those reports were being built much faster. They could be manipulated much faster, distributed and shared. And so after that, we consider that a successful pilot. We then used those advocates as peer resources in the office, going out and helping the other teams to adopt Power BI, learn the secrets, share what they've learned, simply because that peer-to-peer interaction was more comfortable for people than an IT to business team interaction.

Nathan Settembrini [00:22:53]:
That's great. And so what, you know, on the continuum of, it sounds like do-it-yourself, done-for-you, like where did you guys, 'cause like some of the backend stuff in Power BI, like building data models and writing DAX and these type of things are fairly complex and kind of the opposite of what your, you know, the users you described would want to deal with. And so I imagine you guys kind of fall somewhere in the middle where there's some enablement that happens on your side from the, the data source perspective, and then, you know, people are just connecting to that and building the pie charts and bar charts. Is that a good description?

Tony Gondick [00:23:33]:
Yeah, that's a fair description. We do curate a lot of the tables and the fields to make it understandable. And really, I don't want it to appear that my IT team is aloof and, you know, just turning people loose with all these technologies. We're still there. But with these advocates, what we found is they were more willing to go out and self-teach and self-learn how to use these technologies. And they would only come to us in IT when they encountered a challenge. You know, there's this great tool called Google and YouTube. People would go out there all the time and they would teach themselves what the best practices are in using BI reporting. Then if they couldn't make that connection, if they couldn't get the report to do quite what they wanted, they had already done a lot of the legwork. So then they come to IT somewhere in the middle and say, okay, I can't get it to do this. I can't get it to move any further. Can you help me? And it was a much quicker turnaround in resolving the issue rather than starting at the very beginning and trying to teach and bring them along. In that very lengthy journey. So we found a lot of success in doing that and we just encourage more and more. We now have more advocates that are going out and even without approaching us to say they're doing this, they're just going out and taking the online learning courses and coming back to us with that, that middle of the process question, the things that they can't quite understand.

Nathan Settembrini [00:25:20]:
That's great. Was there anything that surprised you now that the business is out there doing some of their own analytics?

Tony Gondick [00:25:31]:
There's some folks that you wouldn't believe are data nerds, but once they got into the data and they could see how easy they can manipulate the Power BI, the dashboards, and create these stories, these graphics, of the data, they really got into it. There were folks that, you know, you look at them for the first time and you make an assumption on their personality and how they are. But underneath it all, they're data nerds. And they love to be able to investigate and see what our weather patterns doing, how does that impact our product? If I were to look at this particular area of a city and state, what are the buying group, what are they doing? They can tell you all of these interesting stories of what they found diving into the data and having those visuals. Because like I said at the very beginning, data is a story. It's got a story to tell and these people now are those minstrels, those storytellers that are taking this and they're telling the story of the business of their particular team and how we're helping them build, or how they're helping the business build success.

Nathan Settembrini [00:26:55]:
Wow, I love that. Yeah, I mean, that is the dream, right, of analytics is you put the power in the person's hand who can actually, you know, they have the context of the business questions and they actually have levers that they can pull to then take action. But a lot of times historically, those people were unable to self-serve with the data. And so you'd have to ask IT and you get a spreadsheet and now you're like, okay, I gotta do 700 pivot tables and—

Tony Gondick [00:27:30]:
Oh yeah, the inner outer joins. Yeah, the key lookups and everything else. Well, and now what we've also seen along those lines is they're more willing to go to ChatGPT or Perplexity. To Grok and ask specific questions. Now, we did warn them, you know, AI is going to make mistakes. It's only going to be able to tell you what others may have asked and what it may have found on websites. So take what it says with a little bit of skepticism. Don't do 100% of what it says because it could blow up your queries. It could, extend the runtime. But if you have questions, if it just doesn't sound right, then let's talk about that. Let's talk about what you're trying to do and have a more deeper conversation of how IT can help you. But we are seeing more of that self-help because the tools are out there. They're more conversational now. They're becoming smarter to understand context, slang in certain areas, and be able to come back with some really interesting thoughts on how to use the data differently.

Nathan Settembrini [00:28:48]:
That's great. Let's pivot to Agentic AI as a differentiator in your business. So I think most people, when they think about AI in manufacturing, they think chatbots, robots, But you guys are using it a little bit differently. So you have a handful of different agentic AI systems and an early warning system. Could you just talk us through some of that landscape and the impact that it's made?

Tony Gondick [00:29:20]:
Yeah, so much like our internal customers are mechanics and engineers, the same goes for our external customers. As we take a look at the authorized Kawasaki dealerships, these are mom-and-pop shops in a variety of different towns and metro areas. Areas that are servicing the, the zero-turn mowers that our OEMs are building. So they know how to fix that engine. They don't know necessarily what all could be done within a computer system. And so when we made the choice to start implementing these agentic AIs, we started very small because we weren't sure what the adoption was going to be if something was there online that you could talk to and ask it questions of. So we're taking a look at just 5 use cases right now. What's my order status? What is the tracking information on my shipment? How do I reset my password? What's my username? Or where is this part used? These are common questions that come in regularly to our customer care team, and we chose those to help reduce the number of phone calls that those questions might generate. And so, as we get ready to turn these on and start using them in a much larger pilot, we are seeing general upward trends in increases of usage. And so, we'll continue to evaluate that over time. Just to make sure that we're hitting the mark and we're actually producing agentic actions that our customers want. If we start to see a downward trend, obviously that's not achieving the goal or it's not as great as we think it is in an adoption standpoint. So we may remove that and insert something else, but working closely with the teams, we're identifying what are those repetitive questions that we can help automate. Then with the early warning system, this is more of an internal, uh, business team system. We would constantly have different warranty claims and calls coming in about part failures, and in some cases it could lead to a large service campaign, an in-field repair that might be needed. Or heaven forbid, a recall of a particular part. What we felt was if we're reacting to those situations, then it is a far more expensive exercise in doing the research to find out what happened, assembling the team to deal with that, the messaging, the customer interfacing, managing all of what's been fixed, what still needs to be fixed. And so we felt there was a better way of taking that unstructured data and building a more proactive system. And we've been able to do that now using Anthropic as an underlying system there where we can take in our transcribed phone calls, where we can take in our warranty claims and our trouble tickets. Our customer sentiment scoring, and we're, we're taking all of that unstructured data and building a central repository that Anthropic can then go out and use business rules to say, we're seeing an increase or a trend in this area. We recommend that you go and you monitor this situation or you do the investigation, because if we can get a ahead of that scenario, we can then build the team in real time. We can identify the messaging. We can have an idea of who is all impacted in the scope of that impact, and we can address it much faster before it could ruin our brand reputation. And by doing that, the overall cost then is much more contained because we control the scenario. Rather than react to the scenario. And the costs then are greatly reduced from a business expense perspective.

Nathan Settembrini [00:33:58]:
Wow. That is, yeah, the early warning thing. I feel like that is the perfect use case for AI today. Like, so my first, in my first job out of college, I worked for a manufacturer, a German company. Kitchen and bath fixtures, and we would import them into the US. And part of my role was after-sales support of these products, right?

Tony Gondick [00:34:26]:
Yeah.

Nathan Settembrini [00:34:27]:
And we would get technical calls and, you know, from plumbers or whoever, installers, saying, oh, this is broken, or, you know, we see this defect, this toilet doesn't flush very well. And, you know, it was kind of up to us to kind of keep track of that. And so one of my early projects was to work with Germany to create a database to track all of these issues, get the serial number, you know, but it was all very human, right?

Tony Gondick [00:34:57]:
And it was subjective. Yes. And yeah.

Nathan Settembrini [00:34:59]:
And someone would have to look at it to try to spot the trends and see like, oh, well, this particular model has this problem multiple times. The serial number seems to go back to this particular mold. And they have to like do that. Rationalization and root cause analysis. And nowadays you just have AI doing that for you. That's amazing. How does it serve it up? How does it inform you?

Tony Gondick [00:35:26]:
So we actually have built it onto Power BI. So we have this nice dashboard that shows all of the different causal parts that it has identified. And you have different statuses where it is Alerting, in investigation, in remediation, or closed. And so you can very easily tab between the different statuses and find, has anything new come up? What is currently being investigated? What's currently being worked? What have we closed in the past? Because that past information is still another data point. It's another story. So if we can add that on top of anything new, it gives us a more well-rounded view of all of that information. But, you know, something that you had mentioned with, you know, being subjective and having a human try to interpret all of those hundreds or thousands of rows of information that are coming in from all of these different sources, you could have the same causal part and give it to 15 different technologists or repair shops, and you're gonna get 15 different explanations as to what happened. What you need is something that can understand the context of what is meant by that unstructured data, because there's a huge difference between an engine misfiring and an engine causing a fire, right? If it's a human, they may not understand the context of the sentence structure or the information that's contained within it. There might be information missing from a description, but AI doesn't have that bias. It doesn't have that partisanship. It takes exactly what is written as cold, hard fact And now you layer on phone calls and claims and case studies and other data sources, and that starts to round out that story so that you can find it actually is a causal part. We are having a failure in this area. Is it the same supplier? Is it in our manufacturing process? How do we improve our overall capacity with this particular part so that our, our brands continue to operate with a high level of quality that they're known for.

Nathan Settembrini [00:38:04]:
Yeah. Yeah. So along the same lines around people, you said that the hardest part of technology isn't the technology, it's people, right?

Tony Gondick [00:38:15]:
It's the people. Yes.

Nathan Settembrini [00:38:16]:
You hinted at this in the beginning. I'm just curious, how do you How do you drive change? And yeah, when, you know, some people have been there forever, some people don't see the value in the change. Some people are very protective of the brand and don't want to change because if we change, we might mess it up, right? So how do you handle that?

Tony Gondick [00:38:40]:
So what we found to be successful is the one-on-one interactions that we have with a lot of the business teams. I mentioned earlier, we find those technology advocates. Those are the individuals individuals on a business team that are more willing to try and be flexible with technology. They're not ingrained in their process. Once we do that, we start to see this, this organic growth of acceptance because those advocates are now talking to their peers and they're excited about what this technology can bring. The time savings that are out there. And so as that excitement grows, we get more and more advocates for a solution. For those that may still have difficulty in accepting this, we do work with them one-on-one. We sit down with them in a conference room. It is just us and them. It's like, okay, here's— what we're hearing. Here's what we think is a good solution. What concerns do you have? If you had the ability to make one change, what would that be? Invariably, they're coming up with ideas of how it can work better within their mindset. And because we're using these cloud technologies, because we're using Salesforce, we can sit there and we can make changes on the fly, have them refresh and say, Did we hit the mark here? Is this what you were envisioning? So we get them to talk more in depth of what their vision is for this, and we try to meet them in the middle. Sometimes we hit it, sometimes we don't. Sometimes it's a little bit more complicated than just rearranging a web page or inserting a new field or a new calculation. So we will have to take that back do the development and then have them test. But we feel if we can get them to take partial ownership over this change, they're much more willing to adapt to that new whatever, new technology, new process, new tool, and then they're encouraged to spread that to all of their peer groups and grow that.

Nathan Settembrini [00:41:08]:
That's great. Yeah, so as we start to land the plane here, I did want to call out the fact that you were awarded the first Golden Hoodie for Manufacturing Cloud ever for kind of a combination of your, like, the innovation that you guys do at Kawasaki Engines, but then also your personal contributions towards the community and giving back and helping others. Is there anything— what was that experience like? And is there anything you'd like to elaborate about the why?

Tony Gondick [00:41:42]:
Yeah. So, it did take me by surprise. A huge surprise. I think it was one of the few times I was rendered speechless.

Nathan Settembrini [00:41:53]:
Oh, wow.

Tony Gondick [00:41:54]:
So, yeah. So, frequently, I'm asked to do different summits or keynotes or presentations. In the manufacturing industry and in the manufacturing space. And so the group that had awarded me the Golden Hoodie had started the same way as it always goes. Are you going to Dreamforce? We'd love for you to speak on AI and manufacturing at the keynote, maybe be about 5 minutes. Well, and then of course I agreed to that. Then, Usually what happens is I will ask for the slide deck of all of the topics that are being talked about in that session so that I can make sure that my talking points match up to the bullets and that I'm not deviating off on some weird tangent during this particular session. But I never received the slide deck. They kept delaying and pushing off and outright refusing to give it to me. So I thought it was kind of weird. Then the day of the keynote, I'm getting flooded with text messages. Are you coming to the keynote? We've got a space saved for you. You have to be here at a certain time. You've got to get mic'd up. Where are you? And they were just getting more frantic as the day went on, right? Yeah. Yeah. We're going to be totally— come on. So They ended— I ended up there plenty of time, 15 minutes to spare, and I'm sitting in the front, but I'm going through in my head all of the bullet points, everything I wanted to say, the people that I wanted to call out, the different technologies. I, I didn't hear a word of what was being said in the keynote. So they call me up there and ask me to introduce myself. I'm on the stage, and as I tend to do, I'm looking left and right. I'm looking at people in the eye, I'm trying to make connections in the audience, and I just happened to turn towards the comfort monitor that was there, and I saw the next slide which said Golden Hoodie Award winner. Well, that completely derailed everything that I was saying, and I, I pretty much was stammering up on the stage, which was the cue for the MC of the event to bring that up and to save me essentially. So you can still see the video on Salesforce+ for the keynote and you can see me just eyes wide as this huge blessing, and I do consider it a huge blessing, was presented because that's who I am. I love helping people. And so if I can help you overcome a challenge that we may have encountered in the past, I consider that a win. I truly believe that whether you're talking data, whether you're talking technology or AI, everyone is starting at a certain place in their journey and someone has been there before. If we can help each other out, we'll make the world a little bit better, and maybe we'll see greater adoption of that technology with all of these great lessons learned.

Nathan Settembrini [00:45:24]:
I love that. I love that. So what's next for you and the Engines division in coming years?

Tony Gondick [00:45:36]:
So we're going to continue to add more data. I mean, that's just a fact of any business. As we continue to build more products and add more data, there's going to be slight changes based on business opportunities, and that's going to change our formulation and what kind of tools we might need to understand that data. So what I envision over the next couple of months this calendar year and really leading into year 3 and 5 coming up is adding more data into our AI models, identifying other business processes like image analysis. There's a huge avenue there that we could leverage AI in, whether it be a manufacturing line or in our testing. Image analysis can find all of the nuances of that test, of the difference between an accurate product and something that might have some variance in how it was built. So we think we can save some time there. But I'm also seeing where I believe we're going to see a consolidation of the AI platforms. It seems like every day there's a new AI platform that has a unique skillset.

Nathan Settembrini [00:47:00]:
Yeah.

Tony Gondick [00:47:00]:
You have, you know, ChatGPT, which is like the Google of everything. They've thrown the kitchen sink into AI and they've got a little bit of everything. Or you have Anthropic, which is really good at document scanning, ingestion, and bringing back understanding. You have Groq, which is really good at heavy data analysis. So I think we'll see a lot of those smaller platforms start to be gobbled up and you'll have primary players with a very unique skillset. And so as an organization, you'll be able to choose a specific AI platform to meet your data needs, your data structure, but also your business opportunities only because of that consolidation. So I, I, it's gonna be an exciting time, I think, in the next 3 to 5 years, and it'll be a lot easier to make choices then once that consolidation occurs.

Nathan Settembrini [00:48:04]:
Yeah, I like that you're doing kind of a hybrid approach between— so Salesforce is a core part of your technology platform, and they're doing a lot of their own innovation and research and development and baking AI into Salesforce. And you guys are making great use of that. And— but you're not just relying on Salesforce to do everything for you. You've also gone and built this, you know, this model that's analyzing all your after-sales support stuff and early warning system. And you're talking about image processing and, you know, vision kind of stuff. And so you're not just relying on the big behemoth, you're also doing some development yourself and kind of looking outside that. And I love that you're not getting too too blindered, myopic with it.

Tony Gondick [00:49:01]:
Yeah, and in the past, we've tried to hammer that square peg into a round hole, and it just did not work out very well. It damaged a lot more than it fixed. And so we made the choice years ago, let's choose the best in breed, the one that was the most flexible solution for a lot of these needs, and continue that over the years, because it was only, what, 4 or 5 years ago that OpenAI turned on ChatGPT and then it exploded. You had all of these other entities, other AIs that are out there. So we want to make sure that we can remain flexible because, as I said before, I've got a small team. We've got to be agile and nimble. I can't spend a lot of resources on learning another tool. So we have to be able to use and reuse a lot of these out-of-box solutions just to keep up with the rate that business is changing. And the more that you can do that, I think the better off and maybe the less gray hair you'll have overall.

Nathan Settembrini [00:50:17]:
Yeah. Yeah. So if you could snap your fingers, and fix one data problem instantly, what would that be?

Tony Gondick [00:50:26]:
I think the biggest thing is the bias that people introduce into the datasets unintentionally, right? The more that you're dealing with data or the more people that you have dealing with data, invariably they can be the most impartial person, but they're still introducing their bias into that data as they collect it, as they condition it, as they provide it to other people. It just happens because we've kind of done this trend through this podcast. Data is a story and they're trying to get the data to tell their story. So if we can remove that bias, if we can remove that impartiality and have the data just be the data, the data will always tell us what the status of the system is or what the state of our business is. We don't need to massage it. We don't need to encourage it in any way. We don't need to remove datasets to tell our version of this story. Let the data be the data and tell the story that it has acquired over the time and with the datasets available to it.

Nathan Settembrini [00:51:54]:
Amen. Tony, thank you so much for being on the podcast. If our listeners want to either follow you or connect with you, what's the best way for them to do that?

Tony Gondick [00:52:05]:
I'm on LinkedIn. You can find me through Anthony Gondik. You'll see my generated AI visions out there. So just take a look at that. I am on X, but mainly from a research perspective. I don't generally post a lot, but you'll also find me kind of wandering around Dreamforce in San Francisco every year and also the different Salesforce events. Looking forward to Connections and TDX later. Summits, I'm usually at those wandering around. So hit me up if you got questions. I'm always happy to grow my community and to help answer questions as best as I'm able.

Nathan Settembrini [00:52:50]:
Awesome. Well, thank you, Tony. And to all of our listeners out there, you know, keep digging and you'll find the truth buried in your data. That's the pod. Have a good one.