Kelly Kohlleffel (00:00)

Hi folks, welcome into the Fivetran Data Podcast. I'm Kelly Kohlleffel, your host. On this show, we dive into conversations with leading voices from across the data community. We explore everything from AI and machine learning to enterprise data strategy, analytics, and everything beyond. Today, I'm really excited to welcome Johnathan Tate to the podcast. He is a seasoned data and AI leader with over 26 years of experience transforming some of the world's most recognized companies.

He's currently the Chief Data Officer at Bridgestone Americas. He's also held senior roles at Deloitte, Nike, Walmart, and Cat Financial. Throughout his career, Johnathan's led enterprise-wide data transformations, launched customer 360 initiatives, and delivered millions in ROI through automation. He's an expert in data integration, governance, and modernizing IT infrastructure. Jonathan, it is a pleasure to have you on the show. Welcome in today.

Johnathan Tate (00:55)

Thanks, Kelly, it's great to be here.

Kelly Kohlleffel (00:57)

Well, let's let's get to it and kind of start at the top. Bridgestone Americas would love to hear more about your role and kind of what your journey, personal journey has been through data over the last several years.

Johnathan Tate (01:11)

Absolutely. Well, as you said, I'm the Chief Date Officer at Bridgestone, and that covers the West. So that's the North and South America regions as well as our EMEA region, Europe, Middle East, and Africa. So it's pretty broad scope. I've got teammates all over the world. We have a pretty good presence in Nashville. We also have remote teammates that work all throughout the different countries and another central office in Europe. 

But we cover in our organization, everything from API and integration work, automation, data platforms, our SAP data platforms, as well as BI and data science, more traditional analytics, and our master data and data governance arm as well. So we've got a great team and really great leaders that make up the organization.

Kelly Kohlleffel (01:59)

It's a tremendous remit that you have. I mean, I'm just looking over that list and I'm trying to go back, you know, maybe even just a decade. And I feel like, you know, some of these areas were there maybe a decade ago. Some of the things maybe a little bit newer. You mentioned data science. You mentioned some of the API integration. Some of that is as evolved as well. Certainly SAP has been around a long time, but how has that profile of functional areas, how's that changed for you over the last, say, five to 10 years?

Johnathan Tate (02:31)

Yeah, it's a great question. It's changed a lot. I still remember the day I was a senior enterprise architect at Walmart, and I had a couple of leaders who believed in me sit down and they say, we want to make you a senior manager of data, data strategy, and we have no idea what that means. So it's your job. So that was probably early in the, I'd say late 2000s when there was still a big focus on BI kind of separately from big data, data analytics. It really wasn't a thing yet.

So I started that journey really investing time and bringing in partners to build a data strategy around master data, strategic data assets that were super important for the company, and kept our focus there so we didn't try to boil the ocean. But then, you know we saw that evolve to big data becoming the next big thing. And I remember reading an article that big data was the new oil. So there's been so many evolving components to that as then it went into leveraging cloud and leveraging analytics and advanced analytics and going from bar charts to contextual data that actually tells a story for our business leaders. They know what in the world they're dealing with, what their business looks like day to day. 

And then the next leapfrog I saw was really in that data science piece where we were using a component of AI machine learning to start looking at historical data and trying to help us predict what was gonna happen or predict what factors we may need to pull or levers we need to pull to help our strategy or our product growth or whatever we were trying to accomplish. 

And then I continued to see that evolve into then telling us things about data that we couldn't even know to look for. I think back in the day, there was a statement that the average data scientist could probably find nine, up to nine, disparate factors related to what was changing the market or changing a history of data. But then you introduce machine learning and AI and suddenly it's infinite, right?

I got to see that evolve and see it kind of explode and it kept me on my toes because as I was leading teams, I couldn't be the leader that just sat back and told people what to do or tried to give them just direction only. I had to do a lot of training myself to stay up to speed on those new capabilities to understand what data science meant when I did my data science certification or even start learning AI and genAI and engineering.

So it's kept me busy for sure, but it definitely helps me as I work with people who are far smarter than me that I hire and bring into the room to come together and create a vision and strategy around the work we do so that we can accomplish our business goals.

Kelly Kohlleffel (05:06)

You know, I always think, Jonathan, the moment that we become complacent and say, hey, I've got this down. I know this. I mean, this space keeps you on your toes. You've got to continually be pushing yourself. What is that next thing? Like you said, where is this going to add value in my organization? Could it add value? To who? And it's just an it's an incredible challenge, incredible responsibility as well when you're running an organization the size that you are.

Johnathan Tate (05:34)

Absolutely. And I'd like to make a comment on that value statement you made. You know, I got to experience that early on journey where we would ask for $2 million in two years to do a master data strategy or a data governance strategy or a warehouse strategy. And what I found is, you know, that really wore out the patience of our business leaders. Our executives would come and say, okay, you know what, you're going to have to start showing some value for this. 

So it shifted my mindset from doing data strategy for the sake of or because I got geeky and knew that it was important to clean our data to thinking and acting like an owner of the business. Why would I do this in the first place? What's the value that it delivers? Is it enough value? Maybe we shouldn't do that part of the strategy because there's other areas we should focus our investments. So really, I think taking that turn from being a data geek, being an analytics geek to how do I think like I own this company and why would I do this data strategy to begin with has really made a huge difference in our team's success and connecting with our business partners.

Kelly Kohlleffel (06:35)

Yeah, it's fun to think about the how. You know, as technologists, the how is really cool because you've got open palette, a million different ways to do something. But you're right. As soon as that how starts elongating the cycle to deliver value, the patience goes down, the relevancy goes down, I like to say, for those executives as well, whoever has, whoever's responsible for it.

Every day that goes by, the relevancy in the organization goes down, and ultimately, folks find a different way around. So I love that philosophy of being very aware that patience is short today in the data game.

Let's talk a little bit about data shaping innovations. How do you, know, especially as you're interacting with executives within the organization, aligning that data value, the innovation that's possible with data, with company goals, and ultimately influencing the future of the business. Tell me how you think about that, how you go about it, and what's most important to keep in mind.

Johnathan Tate (07:35)

I love that question. Ive got a few thoughts. First, it's really, really wonderful that our executives now understand data is so important. We don't have to tell that story anymore. I think any business leader you talk to, they will say data is critical. We have to have the right data. Bad data in, bad data out, we get it. But then what do we do with it? And I think that's on us as data leaders to come in and understand our organization, understand our business processes.

What we do, get hands on, go touch the products, go visit patients if you're in a healthcare setting or maybe the doctors and the nurses that are in that setting. Go out and visit stores if you're in a retail space. Get to know our business and understand what they do because you can't sit back and say, this is data analytics, come involve us, come invite us to the meeting, make sure we're there, you didn't involve us. That kind of mindset is going to keep value from being delivered to the business. But if you go out and learn the business, learn what they do, and understand what you can do with data and analytics to help drive value is gonna be key, because you become that magic person, right? That translator between the technical work or the data work, the data science into here's how we can use it to reduce our fraud claims inaccuracies. Or here's how we can use it to better create polymers within minutes instead of years to create a product. Those are the things I think the business relies on for a data professional and a data leader to understand so they come to the business with those value add statements and ideas.

Kelly Kohlleffel (9:06)

That is such a great point. I think it also, you think about data engineers today, application developers, we focus, I think so often on what are your technical skills? How well do you know this technology? And you talked about, can I have a conversation about fraud claims? Nothing to do with technology, but like the business process around that, or being able to develop polymers more effectively or efficiently, I think that you said. When you're interviewing, and you're adding a new team member to your organization, I would think that this is probably one of the things right at the top of the list, maybe even overrides sometimes the technology aspect.

Johnathan Tate (09:47)

Absolutely. Yeah, I can say so much about talent and the individual. You know, there's really a genius that lies in each one of us, and it's a leader's job to empower and unlock that genius that is in that person and give them the right, you know, equipment, of course, make sure they are technically competent, but ultimately give them the direction and then the freedom to go try and to do and to fail fast and to learn from those failures and come back because with data and AI, there's so much unknown out there still. Yes, we've got better classes, we've got better training, but it's still kind of a get into the garage and see what works and see what doesn't work. And so as a leader, it's really important to create that freedom, but it's also important to find individuals who can operate in that type of mentality, who are able to deal with a little bit of ambiguity. But I would say the single defining attribute that I look at in an individual, because I went through this myself, is someone who wants to know the so what, I call it, or the why.

You know, if we're asked to work on an algorithm or a large language model or build some type of AI capability, I love to find people who understand why we're doing it or ask the question why we're doing it. How does this help the business? What's the value that it delivers to the customer? Because they're connecting the work they do all the way to the endpoint. I will say that's rarer than I wish it was in the talent market. Oftentimes, we have to kind of push our team members to learn that and to behave that way. I've seen really high performers will have that mentality, and they'll ask for the so what so they understand the value and then they pair that value with the work they're doing to drive innovation and honestly to drive just excitement around doing it.

Kelly Kohlleffel (11:24)

That is fantastic guidance. And I would imagine that when you're talking about some of these newer technologies, how they play into the overall organization and the value associated with AI, ML, genAI, and your team is looking for ways to help get it going and make it stick. This guidance that you gave about looking at the so what, asking the why, gotta be right there at the top of the list.

Johnathan Tate (11:51)

That's right, absolutely.

Kelly Kohlleffel (11:53)

Well, developing a team in, let's say I'm interested in, I like what I'm hearing, Johnathan, I'd love to be a part of the organization. What should I be doing right now? I've already heard a couple of things, be very focused on value, focused on the business, ask the so what, ask the why. What else can I be doing as a data professional today to set myself up for future success, whether it's within your organization or trying to have a, maybe an outsized impact within the organization I'm in today?

Johnathan Tate (12:23)

Great question. I love that. I've got a few points to talk through. But first, I would say going back to tools, make sure you're learning the latest that's out there, at least a little bit about it, even if it's not directly related to the job you do today. I've had data scientists come up to me at networking events or community events. And they would say, I love your company. I've heard you've got a great team. I'd love to come work for you. I'm a data scientist. And I say, fantastic. What primary language do you develop in?

And they kind of draw a blank. And I said, what tool do you develop mostly in? And they draw blanks. So just there's an awareness as leaders that some folks are out there selling a story or experience that they don't have. And I would really encourage the individual to grow your skill set, learn it before you say you know it, broaden that skill set, start learning Python if you don't know Python, start playing with the genAI, prompting tools that are out there.

Go beyond just chat GPT-4.0 and try all the others and see how they compare. Then start taking some fun ideas. I mean, there's lots of online communities that do these events where they practice creating a dashboard with AI, where they practice creating a small data science project. I would certainly encourage doing that as well. That way you build your community, you build a group of people that you can go bounce ideas off of, and then ultimately get engaged with your local community and technology group so that you can network and kind of get to know who's who, know what leader you're going to work for, what culture they create, because at some point in your career, culture is going to become more important than money. And you want to know what kind of culture you thrive in and what type of team you want to be a part of. And that's something I work very hard for my team so that they feel encouraged, motivated, empowered when they come into work, and it's something they look forward to versus waking up, hitting the alarm clock and thinking, okay, I got to go into the office today.

Kelly Kohlleffel (14:16)

Having that innate curiosity. I want somebody that's trying the latest tools. Is there anything else when you're hiring specifically, especially in this time that we're in right now, this genAI time, if you will, anything else that you're looking for from a hiring perspective?

Johnathan Tate (14:33)

Gosh, there's probably so many things that I'm not thinking about. But if I was a data professional, you know, starting out, the first thing I would say is go find an advanced class in data science and AI and go ahead and take it. Even if you're done with college, you've graduated, you've got your degrees. A lot of local colleges, especially in Nashville and in Los Angeles and Seattle, the metro areas will have classes that you can take online. You can take part time to get some certifications. It's not about the certification even, it's really just about learning the new skill sets and the new tools. And so I would say to them, go take an advanced course in data science, get that under your belt, do some AI accreditation work. You can get free classes. IBM had a great free offering. Don't worry, I'm not sponsored or paid by IBM. It's just something they have for free. But that's definitely something I look for because you've got to continue to hone that craft and to stay relevant. And then once you get it and you understand it, be ready to forget it all and just do what works for the business and for the team that you're in. 

Having that flexibility to say I'll wear multiple hats or I'll try something else, I'll go outside of my job description for the better good of the team, that is really valuable to have. Kind of going back to my story because we are in an environment where it's constantly changing and now it's more important.

For us to say, sure, we'll try something different or learn something new than it is to say, here's what I develop in, here's what I'm accredited to. So that's one aspect. The other aspect I would say is networking and understanding the business and growing in your knowledge of that industry that you would potentially be working in, learning the company itself, the history of the company, where they've come from, where they're headed. Those are things that I think more and more are important in someone's resume, if you will, when they're interviewing for the job. Because it shows that they're interested in the greater success of the company as well as their own personal professional career. And that's very valuable when I talk to folks and they show an interest because it tells me that they're vested in knowing what we do and why we do it when they come into the role.

Kelly Kohlleffel (16:48)

That's great advice. That's great guidance. Well, let's transition. I love that discussion, by the way. Let's transition into the CDO role in particular. It's rare, I think, these days to find CDOs that have long, long tenure. I mean, this industry is moving so fast. I think average tenure, somewhere in that three-year range.

So you have a limited time window to really have an impact. In your opinion, what should CDOs prioritize in that three years to deliver real impact to the business?

Johnathan Tate (17:21)

That's a great question. absolutely love this question. I think I look at a 30, 60, 90-day path for CDO, and I say really your first 30, 60, 90 days should be an accelerated attempt at building those relationships very quickly with your business leaders and stakeholders, understanding what they do and what their challenges are and then coming back at the end of that 90-day period and saying here's where we can start to add value and deliver. That may require some change. That may require a longer roadmap.

But if you can bring back some quick wins that then build trust with that leader and with your peers, then you'll get the support to build that longer roadmap and to invest in that longer strategy. I think that's critical for any leader. I would say one model does not always fit the box. So there's been lots of conversations around, you centralize data analytics? Do you decentralize? Do you federate? Well, as you learn that business and as you learn that business culture, you'll start to learn whether or not those models fit.

Right, and I've been at companies where centralization absolutely worked, that's how they operated. They gave a leader the authority and power and then they'd say go hammer it down at the organization. Okay, well you can centralize all the resources and the capability and then you can deliver that service out to the company. I've also been at companies where it was a little more collegiate, where everyone worked through a matrix. It was really about empowering each other and helping each other, but there was no clear decision-maker sometimes. And that's where you know, maybe a federated or center of competency model. I like to call it they work where you build a capability that's lean and efficient, but is the expert resource around data analytics and AI. And then you partner them with those business citizen data scientists, citizen BI developers, to create that expertise out in the business. And so long story short, be very flexible and make sure you fit the strategy and the vision that you have for data analytics to the organization that you're going into is really the first thing I would say. 

And second would be, you know, build relationships. And this is getting a little personal for me, but, you know, really think about why you're a leader. And I think for me, that's because I want to lay down my life for those around me and I want to help them become better tomorrow than they were yesterday. So to do that, it's really not my agenda. It's not my ego. It's not what I want to do. It's what do we want to do? How do we create something together?

How do we bring the business leaders and peers along the journey with us to create what the future should look like? I've gotten a lot of feedback from business stakeholders that they value that so much because instead of us telling them what they need to do, they feel like we are listening for the first time and actually being a partner in creating that ultimate solution. And yeah, they know we've got the expertise in the background, but they very much want to be a part of that solutioning together.

Kelly Kohlleffel (20:10)

Wow, that's fantastic. I do want to ask you though, you talked about this 30, 60, 90. How quick do those wins need to be when you say, hey, here's where we want to start?

What's the guidance to your leaders that are out there in the business and they give guidance to those business stakeholders? We will deliver this in…how quick is that when?

Johnathan Tate (20:32)

I think you should have something that you try to develop within the first 3 to 6 months after that 90-day kind of vision so that you're capturing value very quickly. And that can be simple projects. I'll give an example without saying the company name. There was a project to go look at historical claims data. And the leader of that business said, I know my team cannot process all of the claims, so I know they're leaving money on the table. How do I get better at processing the right claims?

So we went in and we said, a data science project where we analyze the last three years of data to understand what we should have analyzed and what we shouldn't have gives us some kind of capability that we then put AI on top of the auto decision claims to a threshold and then carved out the rest of those claims that were very valuable, that left a lot of money on the table. This is what the team should be working on. Okay, that's a quick win. Now the business sees the value.

And you can start small as you build out the rest of your strategy or as you just create trust with the business that you can continue to deliver capabilities like that. I'd say definitely have something within that first 3 to 6 months after your 90-day period. If it's a very, very large company, try to have something within the first year. But don't be afraid to also come back and say, you know, we need to do a data governance program. That's might take 18 months or longer, but have the quick wins ready so that you continue to deliver values you get there.

Kelly Kohlleffel (21:58)

I think it's interesting. talked about what sounds to me, claims processing, that is typically a very core business process. And I think my mentality would be, if I'm brand new, I kind of want to stay away from that. That's probably locked in stone. But it sounds like there are examples where even maybe a core business process, there could be some opportunities for that quick, potentially low-risk type of win that you can get even around something that maybe has been established for a long time.

Johnathan Tate (22:26)

Really that's where you as the leader or the data analytics professional have to start asking those questions and get out to the plants and get out to the mills and get out to the retail stores or the facilities or whatever your company does so that you can ask that critical question, you know, what do you struggle with? What do you wish you could do? What do you have not done? What do you not have enough data for to decision or to understand? And that's where those quick wins come from.

I'll give one more example of product line. If you're creating a product, you can go talk to the quality folks at the end of that product line and ask them, you know, what do you wish you could do? In one case, we found out they wish they could better predict what the quality of that product was going to be so that they knew if it was going to hit A grade and they could sell it or if it was going to be less than and then have to get rid of it. Well, honestly, that led to then a question of who's decisioning the raw materials and the process that goes into it.

Well, it's somebody that's been here for years. They use their gut. And we probably are overbuilding that product so that we don't miss quality. You put data science and AI on top of that. It can tell you exactly what inputs to put for raw materials. It can ensure that you meet that performance from a quality perspective. And you can save money. I know you can, by reducing the amount of waste that goes into the build. So that's where that motto of think and act like an owner really comes into play.

Because if you're a data professional that thinks like the business owner, you're going to start making those connections.

Kelly Kohlleffel (23:52)

This is such fantastic advice, velocity, quick wins, think like the business owners. Is there anything else when you're talking to a business owner, you're talking to a key stakeholder that you want to, that we should be thinking about to get buy-in on a data project?

Johnathan Tate (24:07)

Honestly, you shouldn't be thinking about a data project. You should be thinking about just the business first and what are your pain points and what do you wish you could do and why haven't you done it? Because the business doesn't know. Sometimes it's a symptom of bad data or not enough data or not the right expertise. Today, there's a lot of confusion around what genAI can do and what data science can do. 

So if you kind of throw any assumptions out the window and not even talk about the data project yet, just understand the problem, understand the need, understand the business case and return that could be, that they're trying to chase after. Then you can go back and talk to the team and use your own experience to say, can we help? In my experience, the answer is going to be yes, so many more times than the actual resources you have. So then it's a whole nother conversation of what work do we do first? But it's a good problem to have.

Kelly Kohlleffel (25:00)

In your opinion, Johnathan, is the approachability of genAI kind of help just move, move mindset, I guess, in general around data science, AI, and ML, say, Hey, I can, I can, you know, find this thing out or get the summary done or whatever it is with genAI. Oh, I think my mind may be open a little bit more to doing something within my organization. So I guess the question is, has the personal approachability of genAI affected what you've seen in being able to deliver projects and not have to convince people quite as much?

Johnathan Tate (25:35)

I think it's helped in some way. I think the masses are getting more comfortable with what AI can do, or maybe even some ideas of what it's capable of, much more so than the last maybe two or three years ago. Where I've also seen an opposite effect is you do also have executives or peers who may be scared of it. They don't want people loading their company data into ChatGPT 4.0 or into any of the other tools that are out there for good reason. You don't want to expose and have a data breach. So I'd say yes and no. If you can come in and say, let's set policies, let's put guardrails around this, let's bring this capability internal so that it's protected and do data privacy by design, security by design, then you'll probably have a much better conversation. And I would say it's vitally important that the chief data officer be very good friends with your legal department, your privacy department, and your security leader. All three of those really are critical to a four-legged stool, if you will, around AI capabilities within a company.

Kelly Kohlleffel (26:39)

Yeah, I agree. And it's just getting more that way. I think that, you we look at the importance, you've hit on it multiple times, the importance of data, all the different aspects of it in that AI flow. You've got so many things now that are dependent on data. The legal aspects, the security aspects, I've got to be confident that my environment, nothing is going to get outside. And I'm not going to mention any platforms or technologies, but I feel like today there is a way to put a much stronger line of demarcation or wall of demarcation around your data and still get value out of some of these newer technologies. And I'm really interested to see how that progresses because I think that's going to make it much easier to have those kind of conversations with the other leaders within the organization that aren't as comfortable yet.

Johnathan Tate (27:34)

The point I would add to that is, you you can have internal resources, take some of these tools and create an internal version of it with Python and with other tools. As you say, I won't go into all of them, but there's also vendors out there who are creating these AI capabilities and expertise that would love to come partner and help you develop an internal facing, you know, kind of restricted or cordoned off environment where you can play with AI. And that's where I see the most excitement is when we can give something where they can just play with it and start tinkering and asking questions. That's where I see a lot of really good use cases for AI come out of the business.

Kelly Kohlleffel (28:11)

Yeah, I agree. Anything that, I mean, you've talked about a lot of great project examples today where you've seen a huge value delivered. Any other projects that come to mind on any of these areas that you wanted to call out, maybe a big impact that happened within the organization and the value associated with it.

Johnathan Tate (28:30)

Yeah, there's plenty. In my experience, I've seen a team take an internal version of the genAI prompting that you see out there with the different providers and create that company's own version of it and then give it to employees because it was restricted. It couldn't access external data. They could upload company information, but it couldn't go outside the walls of that company. So it's protected.

And then you started seeing all kinds of really cool use cases from people automating their scheduling and their day to more advanced where they're asking sales questions, they're asking process questions, inventory questions, and the team is being asked to connect more data internally so they can answer it, almost replacing BI in some cases where the business says, if you can just give me a way to ask what I need to know around my daily sales variance and what products are not shipping and where inventory is backing up, I don't need a report anymore because now it's telling me. So that's one really cool use case, I think, that comes to mind. 

There's others around more of the automation side, using AI to create agents that automate kind of the click and drag and the day-to-day. know, RPA, in my opinion, is kind of dead at this point because AI agents have replaced what RPA can do, and they do it so much better. I can now spin up an agent in one of the most, one of the popular tools that are out there.

And I can have it just watch my screen while I do my daily work and then say, okay, now break that into multiple agents that would handle that on a day-to-day process. It's phenomenal. So that's where I'm saying it may automate some of the work we do, but if you know how to train and develop it, it'll never automate and replace you as the human component. And that is one item I'd love to talk about. There is still a leaning, well, let me rephrase this.

I think executives still expect some human component in governing and checking the AI tools that are done. And there's a hesitancy to allow full automation or full autonomy for these AI agents, because they know that there's things like AI hallucination. They know that it can have a hiccup and suddenly it's got too much access, and it's changing things in production it shouldn't. So I've noticed that when we're able to put that governance in place and still have a developer that checks the agent checks the deployment, the testing, and then promotes the production, we get a lot more support than trying to go in with a business case around, we're just gonna completely automate this process and nobody will ever have to touch it. Just food for thought there.

Kelly Kohlleffel (31:10)

Yeah, I know that's fantastic. Anything that we haven't touched on, just any other data trends, you kind of talked about a downward trend maybe with RPA, right? Anything else that's either kind of trending down in your opinion or trending up. Obviously you talked about genAI, RPA, anything else that you've got your eye on?

Johnathan Tate (31:29)

I think on the BI side, if you can't create an incredible dashboard that tells a lot of information and a lot of stories with that data and context, then you're probably going to be looking at what else you can do. Because the basic charts, the basic answers, the bar charts and line graphs and even dual access charts, that's all getting replaced, I've noticed, by AI agents where you can kind of do that prompt and questioning and response type query.

So that the business has that on demand versus I've got to wait for a report to be built. I've got to wait for that to refresh every night. I think that that's kind of changing the landscape. I wouldn't say BI reporting is out the window. I'd say BI reporting is evolving to be more of an expertise around creating large scale like executive dashboards, a really good story-driven dashboards. And then AI agents are going to be able to handle the day-to-day simplistic reporting that you typically would have a BI developer do.

Kelly Kohlleffel (32:28)

Very good. Very good. 

Johnathan Tate (32:30)

So that's one. I think code, and I would probably say all developers all over the world are going to love this. I, being a former developer that hated it and quickly changed my career's choice, loves the fact that you can use AI agents to write the code for you. Now, you still have to know the code. You still have to understand it. You still need to know Python, for example, and what connecting the libraries will do and all of that because you need to review that code.

In some cases, you may need to tweak it and change it. But the day-to-day trudge of just writing code all day long is quickly going away and being replaced by these AI agents where you can feed it the requirements, it can write the code, and then you simply verify, adjust, and then create the output.

Kelly Kohlleffel (33:14)

Is there any use case where I should just write the code from scratch today? What's a scenario? Because I don't know that I can think of one. I'm not a developer, but why would I want to do that versus accelerating that cycle like you said?

Johnathan Tate (33:29)

Well, if you don't have the AI tools or you’re not allowed to use them. Obviously that's one example, but if you've got the tools at your fingertips, I don't think there's any reason you couldn't. Maybe when you get into the space of modernizing old applications and all that stacks, that may be a little more difficult. And so you might have to break that down into more bite-sized chunks or do some of that manually. But I'm even seeing that evolve to where we have old systems that are still on-prem because they're running on AS400.

Johnathan Tate (45:47)

And now if I've got a really good AI engineer, they can start creating agents that break that code down and rewrite it in more modern languages that then can be cloud hosted. So, you know, maybe a little bit here and there is still going to be always be there, but it's very, very quickly evolving.

Kelly Kohlleffel (34:16)

Yeah, that's a great example, like refactoring old AS400 RPG code. Or I think about also data migrations, like refactoring legacy ETL workflows into something more modern, where this took millions and millions of dollars, just ridiculous amounts of hours to do that. How much can I shorten that timeline down when I'm using AI, genAI and AI agents to do that and probably getting 80, 85%, maybe 90% of the way there very, very quickly.

Johnathan Tate (34:52)

Yeah, absolutely. It's going to be a game changer.

Kelly Kohlleffel (34:56)

Well, Jonathan, this has been a just a wealth of insights, a lot of fun. I really appreciate you joining the show today.

Johnathan Tate (35:03)

I'm glad to be here. I always look for opportunities where I can give back what others have invested in me and in my career. So I appreciate you giving me the time to share.

Kelly Kohlleffel (35:12)

Well, thank you so much. Look forward to keeping up with everything that you're doing. I think this show today, by the way, if you're, doesn't matter if you're a data leader, you're an individual contributor. Johnathan has just shared some incredible insights. So thank you for all of that. Really appreciate it.

Johnathan Tate (35:26)

Thank you. Have a great day.

Kelly Kohlleffel (35:27)

Absolutely. And a huge thank you to everyone who listened in. We really appreciate each one of you. We'd encourage you to subscribe to the podcast on any of the major platforms, Spotify, Apple, and Google. You can also find us on YouTube and visit us at Fivetran.com/podcast. Please send us any feedback or comments@podcast at Fivetran.com. We'd love to hear from you. See you soon. Take care.