Experts of Experience - Why Data-Ready Companies Are Winning at AI

Episode Date: January 28, 2026

What separates companies succeeding with AI from those struggling to see results? The answer isn't the AI itself-it's the data foundation underneath it. In this episode, we explore why the companies winning with AI today are the ones that spent years cleaning, structuring, and curating their data. Our guest is Mike Hutchinson, Chief Operating Officer at Teradata, bringing 35 years of customer experience expertise across three software companies. Mike breaks down the critical gap between structured and unstructured data, why vector stores and context engines are game-changers, and how real AI use cases come from solving business problems-not from hackathons. He shares examples from banking, airlines, and Teradata's own operations, including their fully autonomous account planning system. We also look ahead to 2026 and Mike's bold prediction: this will be the year companies build agents at scale - autonomous systems that monitor, decide, and act without human intervention. Chapters 00:00:00 Meet Mike Hutchinson, 35 Year CX Veteran 00:09:22 The Unstructured Data Problem 00:12:07 Marrying Structured & Unstructured Data 00:16:15 Why AI Use Cases Fail 00:19:32 Banking NPS Transformation Story 00:25:18 Security & Hallucination Concerns 00:30:32 Getting Started with Data Organization 00:37:33 Predictive Analytics & Weather Patterns Example 00:42:21 CX Skills for the Next Decade 00:48:52 The Expertise Debate in an AI World 00:56:49 2026 Prediction: The Year of Agents Watch Next: https://www.youtube.com/watch?v=enBE_5PQOb4 Experts of Experience is produced by the team at Mission.org.

Talked a lot on the show about AI use cases and some that have succeeded and some that have flopped and like sort of the over promise of AI. And I'm wondering if you would say that maybe one of the biggest gaps is this data problem. I think there's two things. Yes, there is a data problem. While you may be able to use AI for some really cool things, if you don't have that data, you're going to end up with AI that's not going to be really valuable to you. The other case that I'm seeing is really the approach to get to them. the best use cases are the ones that are starting to solve a business problem. You really need to start out with what's the business problem I'm trying to solve and then start from there and back into a use case.

And then, you know, if you've got the data ready to go, you're going to be in much better shape. Mike, welcome to experts of experience. Hey, thanks for having me, Lacey. Glad to be here. Of course. Of course. I'm really glad to have you. Before we dive in too far into what we're going to talk about today, I would love to just get

a quick intro from you. Where are you calling in from? what company do you work for and what's your background in CX? Hey, thanks Lacey. So I am talking to you from just a little bit north of Atlanta. That's my home base. I am with Teradata. We are a software company headquartered out in San Diego,

which is where I'd rather be today. But, you know, my whole life has really been in CX. I've spent many years in consulting, running customer success organization, support organizations. So the experiences I have have all been in the customer space since I really started my career way back when. Yeah, that's amazing. And how many years have you been in the space?

Let's see. I've now been 35 years across three different software companies. Like I said, always focused in the customer experience space. What have you seen evolve and change in those 30 plus years? Well, that's a lot. That's a loaded question there. I mean, if you really look at it, you know, as time has gone on every year as we move forward, companies want to know more and more about their customers.

CRM is not something new. I mean, it was started years ago. But what's really changed is the fact that today there is so much more data that we can get about our customers, whether it's in a B-to-C or in a B-to-B situation. You know, the Internet obviously brought a ton. But even moving forward now, just the event of all social media that is available, we can learn so much about our customers. And that is a gold mine for companies. And they are doing their best to try to mine that today.

Where do you see a gap in being able to mine that? Because I know when we chatted before, we talked about this a little bit, and I would just love to set the table for our audience of like, yes, we've got access to all this data. But that doesn't mean we can use it. Yeah. I mean, you know, look, for years, what everyone's been really good about is really kind of. capturing all that structured data. And so, you know, we have details on customers and accounts and, you know, what they purchase with us and their purchasing history.

We've done a really good job of curating that data. And it's available and customers have it locked up in their enterprise data warehouses all over. But what's not being captured today is all of that other information that's available, whether it's from voice or text or video, chats, you know, that's the new frontier. And, you know, the problem with that is, one, it's the volume of that information. There is just so much available to companies today, to mine. But then the next problem is how do you organize it? How do you structure it?

How do you search it? How do you make meaning of it? And that's where it gets really complex. It's not just a volume issue, although that's a part of it, but it's a structure issue. How do you mine all of that unstructured data to get true benefit? How have people historically been doing that in the last couple years? and obviously now that's shifted as we have AI as a tool to help us with it. Yeah, I mean, what I'd say is over the past couple of years,

they haven't been doing it well. You know, it really is, you know, in many cases, what it's trying to do is take some voice or video and, you know, obviously translating that to text, then putting text into some site search engine and, you know, really trying their best to get what they can out of it. But it's a complex problem. And, you know, obviously, you know, we're going to talk about AI today.

And what's making that a lot easier is just the capabilities that AI is bringing to the marketplace. So, yeah, let's just shift right into that with what AI can do now. How are you seeing people not only being able to clean up this unstructured data, but actually use it? Well, it's interesting. I think what I see right now is customers are really trying to tap into it. And so maybe it's tapping into text. that they got or chats that they have with customers. And so being able to bring that in through into a vector store,

which then gives you an inherent capability to search this unstructured text to make meaning of it and to gain insights from it. You know, that's step one. But I think what I'm seeing in the market is people are beginning to fall short because it's not just take chats for an example. It's not just enough to search through those chats and understand what those chats say. now you've got to get real meaning. You've got to get context from them.

And the question is, how are you going to go about doing that? And how are people going about doing that? Is it just with AI or are there other ways? No, I mean, the best thing to do is what you've got to do is you've got to marry up that unstructured data with your structured data. So if you step back and think of an example like a bank, a bank is going to have all that information about all of their customers. It's going to understand, you know, how valuable they are as a customer, you know,

what type of products do they have? What's their annual spend with the bank? Is the bank making, frankly, money off them? That's their job. So the value comes in and taking all of that structured data that you know and then marrying it up with that unstructured data. So think of it in terms of chat. So if you're a bank, you may have thousands upon tens of thousands of chats a day

coming in through your portal or through a mobile app where customers are asking for things, complaining about things, recognizing things, recognizing things. Where AI is really coming in is to take all of that unstructured data, make sense of it. But then to combine it with that structure data, so you get some real meaning. You get some real insights. Yeah.

How are you guys thinking about this at Teradata? So, you know, look, what we know at Teradata is we, you know, we are custodians and our customers. We have that structured data. And we have many of our customers for 30 years have been curating that data. It's well organized. They understand what's in their enterprise data warehouse. Really what that's turning into is really a knowledge platform to marry up with that unstructured data.

So at Teradata, one of the things we're doing is we're offering customers the ability to get to that structure data and combine it with that unstructured data through tools like a vector store, an MCP server, an agent builder. And what that's allowing customers to do is to take that foundation that they have in that enterprise data warehouse and really turn it into a context engine for AI. And we're having great success of that. We've seen really great examples of customers being able to get tremendous value out of something that they've been building for years. But now they're using it for a new use case. And that new use case is, how do I make my AI better? How do I give it context?

Do you see your customers that are really excelling or just in, I guess in general, in the market, are you seeing the people that succeed at these types of AI use cases, the ones that were spending the time years ago, cleaning up the data, thinking about how I'm going to maintain this over time? It's something that I've heard across the business world and across our podcasts over the years about, like, the importance of cleaning up data. But I don't think it's something that people really felt super strongly till the last couple years when they've been trying to implement AI solutions and not had proper data to use. So I'm just wondering if you're seeing that

where these companies that had started years ago thinking about this problem have a significant advantage today. They absolutely do. And I mean, that's one of the, it's interesting. Not only do we see it. And our customers, just by the nature of the use of our platform, have been doing that. And so we're seeing our customers have a huge leap ahead

when it comes to AI because they have spent the time to analyze that data, have it structured, have it clean, have it accessible. And it's interesting, not only are we seeing it, our partners are seeing it, because suddenly we've got a lot of partners coming to us with the recognition, hey, Teradata, your customers have this data, and they already have it structured, and they already have it accessible, and now with AI we can give it meaning. So how can we help, you help them get real value out of AI?

I'm wondering from your perspective, with, you know, we've talked a lot on this show, show about AI use cases and some that have succeeded and some that have flopped and like sort of the overpromise of AI. And I'm wondering if you would say that maybe one of the biggest gaps is this data problem between companies that are able to really see success with AI and those that are not. Or is there something else that you're seeing as well that's preventing people from really succeeding at AI use cases? Yeah. I think there's two things. I think number one is, yes, there is a data problem. And so while you may be able to use AI for some really cool things, if you don't have that

data available that's going to give AI context, it's going to give AI the guardrails, it's going to keep it from hallucinating, you know, you're going to end up with AI that's not going to be really valuable to you. And so one is, yeah, it is a data problem and the customers that have spent time curating their data and building these data sets and data products, they're going to be ahead. There's no doubt about it. We're seeing them. I think the other case that I'm seeing with use cases, though, is really the approach to get to them. And what I mean by that is, you know, the best use cases are the ones that are starting to solve a business problem.

And that's a much better use case than the one that's come out as part of a hackathon or some kind of sponsored activity where, you know, teams are like, hey, let's figure out what a great use case is for AI. And here it is. you really need to start out with what's the business problem I'm trying to solve and then start from there and back into a use case. And then, you know, if you've got the data ready to go, you're going to be in much better shape. It is interesting because I do think, you know, we talked to Samit Aurora. He's our chief product officer. Yeah, so we talked to Samite Aurora, your chief product officer on IT Visionaries,

one of our sister shows about this. And he'd mentioned the hackathon thing. And I've heard about that a lot. And I feel like I've actually talked to people over the years about this. And I'm like, oh, I'm realizing. maybe this was just part of their AI hackathon whenever we were discussing it, right? But that is an interesting thing that I think every company can feel or has felt. It's just this like, we must use AI.

So I'm going to think of the thousand different ways I could possibly work it into my current work stream. But there's not the pause and zoom out and think, okay, what's the, as you mentioned, the business outcome or the business problem I'm trying to solve for. And then back into how can I solve it with AI? I'm seeing just a lot of people trying to replace steps in their process versus how do we just completely rethink this entire process to get to this certain outcome that we're trying to do. No, I mean, you're right. And, you know, look, hackathons have a purpose. A lot of times they're great for, you know, education and getting people up to speed on new tools and technology.

But to, you know, to use a hackathon say, okay, we're going to come up with our use cases for AI. Not a great approach. Well, and I do like the idea, though, of having those types of things, or at least having that dialogue in your company to encourage imagination and creativity around these new tools. So, yeah, to your point, I think that is like a valid use for them, but whether or not we're actually going to take those AI demos and turn them into something for our companies to be determined. So I'm curious from your perspective, working with all these companies and over the years, have you, what's the top successful AI use cases you've seen come out of this? Yeah, let me give you a couple of good examples. And these are all recent. And I think that the best story behind all of these is they're short term in terms of development and getting to deployment.

And again, it goes back to the fact that, hey, you've got all that use case available. So one of them, I was really kind of talking a little bit about when we were talking earlier. And this is a bank coming out of the APJ region. It's a large multinational bank. NPS is really, really, really important to them. And so, you know, what they they were seeing is they were seeing, you know, a decline in their MPS scores, which was really concerning. And so what they did realize in the example I gave before is they had all this wealth of knowledge

and chats that they never used. And so what they were able to do is working with them, we're able to take all that chat data, get it into a vector store so that it can be searched, married up with that structured data so they can understand, you know, their most valuable customers. what are they saying via chat? Out of that came a number of recommendations, and over a period of three months or so, they implemented them. And as they've gone through their next MPS cycle,

they've seen a significant increase in MPS store. And that's really hard to do in a short term. Wow, yeah. Only three months? Yeah. That's brilliant. And so what they were able to do is with the help of AI, but taking all of that chat data,

is really understanding, frankly, what was irritating their most valuable customers and put in things to rectify it. So that's a great use case. Another use case is a large European airline that we've got. They also were looking to understand customer sentiment. And again, around the customer experience point of view. And coming out of it, what they determined, their biggest problem was their baggage. And how their baggage handling was so negatively.

impacting their overall customer sentiment. And so, you know, I think, you know, they were able then to quickly, you know, start, haven't fully rectified it yet. But they got to a very quick understanding of what it was that was, you know, really impacting their brand. And so really focused on that. You know, so there's lots of good experiences that I've seen with customers, even internally at Teradata, one of the things that we've done is our account

planning is now fully adjunct. And for all of our customers, all of our sales reps have an updated account plan every week. And that account plan takes into account everything we know about a business. It's all of our internal data. And so our data out of Salesforce, our data out of service now, or data out of our telemetry system, which monitors their system. But it also takes into account all of that public information that customer have. And so, you know, as you talked about earlier, sometimes people look at just let's automate it'll, you know, some step in a process. What we've done is we've rolled out a full end-to-end

new account planning process for our go-to-market team. And every A is delivered an account plan on a weekly basis that really is up to date on everything we know and is publicly available about that customer. You know, when we were talking a few weeks ago, you mentioned that you are maybe one of the biggest, you know, users of Teradata and critic, too, because you are a user And you're going to your team and you're like, hey, we need this. And so it is so cool hearing you talk about how you guys are using this internally because you guys really are your first customer. You're using all this data. You're matching it up.

So, yeah, I mean, how do you kind of go about doing that kind of like feedback cycle with your team internally? Yeah. So, I mean, we consider ourselves what we always call ourselves as customer zero. And so, you know, the efforts that my team, you know, our job is, you know, I'm a customer of tarotata from that point of view. And so, you know, what we are giving feedback constantly to submit, who you had talked to earlier, and his team is, hey, here's what we really need. And frankly, if we need it, probably our customers need it too in terms of features and capabilities. We're also really making sure that, you know, when it comes to AI, one of the important things to scale. And so we do things on massive volumes.

So we have, you know, every Teradata system out in the world. We're getting telemetry on that every second. And so, you know, we're able to deal with massive volumes of data as well. So not only do we look at it from kind of a feature function point of view, but we also look at from a practical point of view. And, you know, we provide that feedback both to our product management team and our engineering teams, really almost on a nearly daily basis. And so, you know, I think some meat would say I'm his best customer,

but he'd also say I'm his noisiest customer. And that is fully with intent to be that. That's great. That's great. I was going to say, they probably love and hate that daily know from you. Absolutely, they do. Absolutely. That's great.

Well, we've talked a few about a few examples from these large companies, like a large bank, a large airline. If I'm a smaller organization or a medium-sized business, are these also use cases, or are you seeing companies also execute on these use cases as well, even if they're on this smaller or medium-sized? Because I do, like, budgets are tighter, the amount of investment you can make, maybe the amount of data you even have because you're a younger company. I'm just curious how those companies are faring in this. Yeah, absolutely.

So we've got a number of what we call new logos. And those new logos may be a division of a larger company or maybe a small organization themselves. What they're finding is the value in Teradata, because Teradata can scale up. We can scale up to and have scaled up to the world's largest companies. But we also have a number of smaller organizations or departments even that are using Teradata for their own purposes. And so scaling is something that, you know, we've done for years. And so we can go up to the biggest. We can work with the smallest as well. The other thing that's

important to a lot of our customers is just where they're going to run this. So we still have a number of customers and even new customers that aren't ready to move some of their most valuable and most protected data to the cloud. And so we offer both an on-prem version of Teradata and a cloud version of Teradata across the CSPs. And so it's a flexible tool that allows you to really bring AI to where your data is. So you don't have to think about, oh, I've got to move my data to the cloud. No, you don't. You can bring the AI tool to wherever your data is and do what makes most sense for you as an organization.

From that security lens, are you guys running into any concerns about that besides just like, I don't want my data in the cloud? But, you know, there are, I've heard it so many times now about hallucinations in the data, things being misrepresented. Like, is this something you guys are regularly seen and you're helping clean up or how are you kind of censoring that? Yeah, I mean, look, there's multiple aspects of data. So, yeah, there is an issue with some customers where, you know, for data sovereignty reasons, maybe it's regulatory reasons, or maybe it's just, you know, how sensitive their data is,

they're choosing it not to move into the cloud. I think the way that we're really focused, though, on, on, on solving the hallucination issue is really you've got to put the guardrails on AI. And so being able to tell AI, hey, this is where you're going to get the facts. If anything in this realm of questions come to you, here's going to get the facts. And so for us internally, one of those, for example, is I can't have AI guessing what our total ARR or annual recurring revenue is for a company, right? Yeah.

We have it clearly outlined, and if you were going to in our enterprise data warehouse, you know where to find it. It's a matter of just telling AI the same thing. Hey, this is the definition of ARR. This is where you get this information. We guide AI. We govern AI. We put guardrails around it.

And so you've got work to do. You've got to protect yourself from the hallucinations. And the way you do that is by just making sure that, again, it goes back to having, that knowledge base, you know, and what we call Teradata is the knowledge platform to really give AI that context. Yeah, okay, that's brilliant. That is smart and thinking about how to map that because there is so many times, I mean, I'll use the consumer example of like, I'm on chat GPT and then it just makes something

up randomly, right? And it's just pulling, but if I had said, hey, here's my document, only pull information from this document, it's far more effective at maintaining a consistent response. Absolutely. So, yeah, it is a lot of that. And is part of that process for you guys kind of troubleshooting, like, what information do we need to provide and having some sort of system for flagging, hey, we're noticing that there is some sort of hallucination or something starting to happen here, and we need to fix that? Yeah, absolutely. I mean, one of the things you've got to do with AI is you've got to always, you know, you've got to always look at the outcome and, you know, make sure that every day it's accurate and every day it's getting better. You need to expect that there are going to be hallucinations. But, you know, the more you tell AI where to go get its answers, you know, let AI do the thinking of the different, you know, different problems to solve.

Let it do the thinking of different scenarios. But when it comes to, okay, now let's go solve this scenario, making sure it understands, hey, here's where you need to pull this information from, that's where the key comes from. Yeah. And this is the part of AI that's just not the flashy, like, sexy thing, right? It's like, you've got to go. Yeah, it's like, oh, we still need to clean up this data. We still actually need to map it properly. We need to think about all this, like, where it's coming from, what's accurate.

Oh, actually, all this data from three years ago, maybe we don't need to use that. We only want to use the stuff that's from like two years ago because it's more relevant to now. Like all that happening in the back end is not the like, I made a flashy new AI demo that people want, right? Well, you know, I met with a partner of ours, and one of the things they told me was they were using AI to support their products. And they spent a year and they put in all the documentation around their entire product base. And I said, well, how'd that go? And they said, well, we spent the nine months taking out 70% of it. And again, because, you know, they didn't think how important that unstructured data was in terms of documentation.

and suddenly you're giving AI too much to choose from. And making sure that it can only look at the relevant material is absolutely key. Yeah, that's great. That's great. I'm thinking, too, about companies that, let's say they are younger or they just haven't been managing their data well. Like, it's been years and they haven't been listening to what everyone's been saying for 10 years now to clean up your data, maintain good data integrity, all this stuff.

What do you, like, what advice do you have to a company? like that that's realizing, oh, shoot, we messed up, we don't have this. Is it just like, hey, you need to invest now and get this rolling now? Or, again, if I'm a young company and I just don't have data to pull from, like how am I going to start to accumulate it so I can really compete? Yeah, what I'm going to say is even if you're young company, you probably do have data out there, right? It just may not be dated. It's in your four walls and you're thinking about. But, you know, if you're a consumer company, there's tons of data out there because people are talking about you somewhere. And so, you know, being able to grab that data, I think that the beauty that the

customers have now that they're starting this, though, is they can actually use AI to help them with that problem. Help, you know, use AI to actually sort through the data they do have, help them get it organized, help them understand what's relevant. And so there's a new tool set that, you know, we didn't have 10 years ago when we were all struggling to get our data organized and we were doing it kind of brute force ways through queries and matching and looking at this. Now you can use AI to help you. And so, you know, really, I don't think there's an excuse not to get your data in order. Absolutely.

And what's the response been on the actual customer side? So we've talked about how businesses have been implementing this. Are you seeing any like, oh, you know, I'm going to share the example of the bank, the NPS increasing, but are you seeing more examples of that where the customers are actually like, oh, we're feeling this. We don't understand it. We don't know what's going on the back end, but we are feeling like something's improving here

because of this, you know, AI tool or whatever. Yep. You know, you do hear the antidotal stories from customers where they're having experience. You know, it's really interesting to me as someone was talking the other day about how, you know, they were 10 minutes into a call and it was trying to solve an issue.

In this case, it was, an insurance issue. And there were 10 minutes into a call before they finally realized they were talking to a chat bot. You know, they were talking to an AI agent, but the AI agent was doing such a good job in terms of, you know, understanding what they were calling about, you know, having relevant information. And so we are seeing, you know, we are seeing this improve for the, you know, the ultimate end users, those ultimate customers.

We're seeing to improve month over month. And, you know, the experience. they're not perfect and they won't be perfect for a long time, but they're improving pretty quickly. And you're getting, I see great examples both within our customers that they're providing to their end customers. And I see some great examples personally as well where that's happening. So we spoke to a woman from Qualtrick.

She had thought leadership there and there was like a new survey that they did. And part of the response was that like customers are still hesitant. Like they're still not seen because they, one person might have this great experience with a chat bot and not to be able to tell for 10 minutes that it's a chat or an AI person that they're speaking with. But like 90% of the other people are having awful experiences with a chatbot or whatever. So it is definitely skewing people's response to this negatively, unfortunately, since there are really great examples of it being successful. But I'm optimistic that over the next year or two, it's going to start to even out and we'll start

to see like more positive response. But are you guys kind of seeing that as well? Yeah, I mean, there's always, you know, there's a hesitant seat with AI right now because the last thing you want to do is you don't want to create a worst customer experience. You don't want to create a bad customer experience where AI is giving a hallucination or giving a wrong answer or, you know, you're on a call with an agent that's, you know, not being helpful whatsoever. So they are being hesitant, but what I'm seeing with our customers is they're picking areas where they know they can make a significant difference and they're making. sure that they do that really well. So instead of like a broad-based approach, hey, here's, you know, here's five things that customers call about on a regular basis. Let's make these almost as perfect as we can make them. And that'll improve the experience. Let's not try to be that generalist when AI is not ready to be a generalist. Yeah, that's really good advice.

But it does take a lot of discernment to come up with what is the thing that we're going to focus on that we can actually solve with this. Because I do feel like so much of this has been like, oh, we're in a slash headcount. We're really, you know, save so many dollars in all these different areas. If we just roll it out broadly and then the response and the feedback from customers has not necessarily been the most positive in those instances. Absolutely. I mean, you really, I mean, you know, at this point with the capabilities AI, I mean, you know,

we're looking at it certainly internally, and I know a lot of customers that I've talked to, we're looking at it. How does it make our employees better? You know, how can it make them a better agent? how can it make them a better rep, a better salesperson. We're not looking at it as, okay, this is going to replace our workforce. We're not there yet. Yeah.

And frankly, I don't think most of our customers want that, but they do want, you know, they do are looking for that agent or a rep to be as good as they can be. And AI can definitely help with that. Yeah, I mean, augment the people that you have, support the people that you have. That's absolutely the right way to think about it. And you feel it. I mean, I was speaking to some, you know, We're shipping things right now this time of year, right?

So let's say, no name shipper. But, yeah, I was talking to them. And it was clear that they, you know, something had gone wrong, but they addressed it immediately. And the woman on the phone was using, like, chat history. She clearly had some sort of integration going on in the back end. And it made it so seamless for me to go from one rep to another. It's quickly handled, super smooth, absolutely loved it. And I was like, this is what it's supposed to do.

This is what AI is supposed to do to support this person that I'm speaking with. Absolutely. Flip flip the other side. I had the same similar experience, you know, with concert tickets or something went wrong. And awful. Like, could not connect, they could not connect the dots. They kept blaming it on. We have this new AI system that won't let me do blah, blah, blah. So like, I was like, this is totally different. This is what it's supposed to do on this one side is like augment the person that I'm speaking with. And on this other side, it's becoming an excuse for the sale or for this customer service person not to support me. So it's just definitely getting such a range of experiences. And unfortunately, with customers, like, the way that our brains work over time as humans is like, if I even have this one negative experience with AI, I'm now wrapping everything up into that, right? Absolutely. So it makes me think even all the positive ones are negligible, because I don't, I don't like it. So yeah, it's just like an interesting psychological problem of how do we continue to like accept that there will be companies that do this poorly and that AI will make mistakes or there will be hallucinations. But ultimately, I'm getting better service, you know,

overall with it. Yeah, it's, you know, it's, it's, it's certainly a dilemma for companies. And I think, you know, the problem is you can really hurt your brand pretty quick, you know, with this and this, you know, and that's what, no one has that goal. And I think, you know, making sure that folks are taking the approach to that they're taking the best approach for customer experience. They're not doing it as a cost cutting, getting rid of people at this point in time. You know, things like that may come down the road, right?

You certainly will be able to save some dollars. But around the customer experience point of view, it's how can you make those frontline people that much better, that much more effective, and actually boosts your overall NPS with your end customers. That's what it should be about. So, you know, we've talked a lot about AI and how it impacts the customer in today, like with today.

Let me rephrase that. Sorry, brain's not functioning. We've talked a lot about AI and how it's impacting the customer and how businesses can use it to improve the experience today. But I also think there's new interesting use cases on how we can actually predict what's coming down the line for customers. So how are you seeing companies use their data paired with AI to really have predict what customers are going to need

and then offer experiences to them before something even happens? Yeah. I think we're seeing it a lot, right? And we're seeing it. And this is something we're really seeing across industries. We're certainly seeing it in the travel sector as an example with all the major airlines. We're seeing it in the retail side. I think what, you know, AI is able to do is, of course, everyone's had predictive models in the past.

They've had predictive models for load factors and what's going to sell and what's not. But the problem with a lot of those models is they can only have so many parameters, right? And when do you, where do you cut it off and say, okay, you know, this is really all we can handle. What we're seeing now with AI, though, is that, you know, an unlimited number of parameters to really be thinking about when it comes to those predictions as they're looking forward. And so if you're looking at, you know, in the world of air travel example, you know, you have all that historical in terms of load and when and where on vacations. but now also taking in the seasonality of the weather patterns and what weather patterns are doing and what are they predicted to do. And how is that going to impact if it's a Lanina or an El Nino in terms of how is those weather and patterns going to predict people's desire to travel to certain places? And so I think what AI is able to do is start really being able to pull from multiple different, kind of multiple different areas of the spectrum,

different things that really can affect a customer's ability to sell and or deliver. And it's being able just to have exponentially more models to be able to look at. And what that's then allowing customers to do is just act that much quicker. And so they can have the eventuality of weather patterns. They can have the eventuality of a global conflict. And what does that mean? And they don't have to react when it happens. they can say, okay, we've already done a scenario that says there's a regional conflict here,

what's that going to mean? Or there's a weather and pattern here. What's that going to mean? Or there's a economic issue, you know, tariffs as an example, whatever the case may be. What they're able to do is just have so many models out there today that they couldn't have before. And, you know, have those at their fingertips. And so they're able to react, you know, be prepared to react that much quicker for the eventuality of whatever it may be that's coming about

their way. Yeah, that is so smart. We have another podcast we host with Lawrence Livermore National Lab. And we've talked to them before about their modeling, like weather pattern modeling over the last 30 years and everything that they've done. And I've been fascinated, by the way that they've been able to put together all these analytics and like systems and maps and everything to predict what's going to occur. And it's so cool to hear that we can now.

do this for businesses much faster and easier than it was for even a national lab to do 10 years ago. And it's just very, very cool to hear all this data is like at your fingertips that you can now access and have different tools available to you no matter what's going to happen. Yeah, I mean, it's, look, it's making the scenario planning for businesses, it's so much easier than it. And it'll continue to get easier because all of these different modelings are going to be, you know, they're available today. more of them are going to come about.

And so it's just, you know, really for a business to understand what are the things that can impact us and let's go grab a model that's going to take that into account. Yeah, yeah. And not be limited by, again, like the scope of the model, right? Like I can actually just go get that information. I don't need to limit my imagination because it's too hard to get that data or too hard to compile that information. Definitely opens up the door for more imaginative thought of how does my business really interact

in the world. I'm curious to hear from you, what new skills do you think people need to be having as we get into this next decade of innovation? You know, it's funny. I just, last week, I just did a call with a bunch of our interns at Teradata. And one of the questions was, hey, what are the skill sets that we, you know, what's the skill sets that we need as we come into this world of AI? And I kind of chuckled a little bit because I said, look, you know, I've been around a long time. And, you know, when I started my career, the Internet wasn't even a thing. And so, you know, the skill sets really, you know, as I told them, it's less about the tool and the technology because you're going to learn that.

And guess what? It's going to change. It, you know, what you're going to learn today, you know, in five years that tool and technology is going to be different. But I think there's some fundamentals that are really important. to people. One is, you know, the data science skill is going to continue to be important just to understand, being able to understand the data, understand data lineage, how important that is, being able to get to true understanding of both your structured and unstructured data. So, you know, that's going to be a skill that is out, yes, there's tools that data scientists use every day,

but those will change as well, but understanding, you know, really that this, almost the theology, if you will, of data science and what does that mean? You know, the most important thing that I think, though, is what I've told these group of interns, it's a curiosity. And being able to win, you know, in this world of AI right now, what I'm encouraging them is go try it. Go try different. Go build your own agents.

Go try different things. Learn, you know, that thirst for learning is probably the most important skill that when I look at our interns that I want them to have. We're going to teach them. You know, we're going to teach them the tools of the day. You know, just having that really a solid background and technology is going to be important. Having that curiosity, understanding the, you know, data science is going to be key. You know, if you have that set of skills, you're going to be hugely valuable, not only today, but for many years to come.

And across the board in any business, yeah, because, I mean, those skills are highly relevant, a matter where you go. Absolutely. And I think about that curiosity piece, because I think about that curiosity piece, because I it's not something that I don't think you can teach that, right? Like, I can't teach you to be curious. But it is something I feel like I'm noticing younger people do tend to have, at least the crowd that I, you know, like I think about my younger brother and his friends.

And I'm like, I feel like people are leaning more into being like skeptical or asking questions of things or being curious about the origin of how something works. So I'm hoping it's becoming more of a standard in how people think. Are you seeing that with sort of the younger generation? I am. I'm seeing it, you know, definitely with my kids, right? I mean, they, you know, they want to understand, you know, I guess what I say is they don't accept everything at face value. They want to understand, hey, how do we get here?

How do we get to this answer? What does this mean? And then what I'd also say is they're not afraid. I mean, they're not a bit, you know, afraid to go try things, you know, and, you know, what I see with them is, you know, in AI, for instance. Apple, they're using it, you know, and they're all professionals, and two of them are in the medical field. One's a consultant. I mean, they're all using it every day, and they're trying things that, you know, in their profession may or may not have been tried before, but they're using what's available to them. They're not afraid of it. You know, they're not necessarily 100% accepting of it

out of the gate, but they're going, you know, they really want to understand, you know, the science behind it, the background behind it, the Y is behind it. And once they get comfortable with that, then just look out because they're adopting it and moving quickly. Oh, so quickly. I know. Are you seeing this with even like more established people at the company, people that you've been working with me?

I mean, you've been working in CX for 35 plus years, you said, like, are you seeing people in your peer group also being ready to adopt it or are you seeing kind of mixed responses? Definitely mixed responses. You know, I was in a with a. a group of peers, not folks from Teradata, but some industry folks. And we just happened to be at a roundtable. And, you know, I asked, I asked a question. I said, how are you guys using AI in your work life or your personal life like every day?

And, you know, one of the people said, I'm trying to avoid it, which it just startled me to, you know, that someone would say that. And he's like, I'm not ready to adopt it yet. And so, you know, what, you know, I just, you know, the, you know, as I tell my team internally, you know, it's not the AI is going to replace you, but people that use AI and understand AI and are embracing it, they will replace you for sure. Yeah. And so I would say, you know, there's some mixed results.

There's a lot of folks that are 100% diving in, trying to figure out how does it make them more productive, how does it make them a better employee? How does it make them better on the, you know, the things they're. have to do at home. But, you know, there's, there's, I think there's a percentage of folks out there, you know, that are still a little reticent and like, okay, this might be a fad, you know, it's the people back in the day that said cloud was going to be a fad as well. So, everything's a fad. And then it's not. And all eight, it's not. I need to learn it now. Yeah, yeah. It's, it is interesting. I mean, we, I talked to a few people this year,

specifically that I would ask the power using AI or when's last time you used chat, GPT. I have And I'm like, what? You haven't been using? Like, what are we doing? Why are we talking? Yeah, it's very, very interesting, the like diverse range of opinions, especially from creatives. Because, I mean, I work in, obviously, in podcasting, and writing and scripting. And there's such a mix of people that, like, you overuse it. And you're like, okay, this is not helpful. And then the people that are, you know, augmenting themselves. Like, I use it as my editor. I use it as my brainstormer. and they accept that. And then the people that are just radically against any type of AI use at all.

And I feel that really strongly in the creative space. But I got to imagine when you're talking about data and CX, hopefully people are going to be more open over time to this. And your peer groups will eventually see it from that perspective. I could only help. So I think they should be. I think it's one of those things that, you know, one of the things I encourage people is use it.

you know, if you're reticent to use it to work, you'll get there. Use it in your personal life. And, you know, the example I had, as I met with my financial planner last week, and I put my entire portfolio through AI, and I gave him a report card, and you wasn't real happy. Because, and I told him, I said, look, AI is not already right, but AI is giving you a great, and it's not that great. Yeah.

Yeah. But it, you know, there's so much that it can help you with personally. And I think once people really start understanding the values, whether it's on the creative side or, you know, just on the work side. I mean, they've got to adopt it. I just can't see any other way. What core skills do you think people will continue to need to have? Like, ones that you think, like the entire time you've been in your career,

even like this skill or this handful of skills are really relevant to people. Is there anything that you think will not change that people should still be really investing in themselves? Yeah, I mean, I think the, like you said, a couple of them that, continuous learning, that curiosity, I think that's important. I think the other thing is just the importance, and this is where on the CX side, it does concern me a little bit, you can't underestimate those, the importance of those human relationships. And so, you know, it's not hard to see in some cases when you get an, you know, an email, and it's from, you know, from a vendor or someone that you work with that, oh, they've, they've had

they write their email for me now. And so, you know, you're still going to make sure that you're nurturing that human connection. You know, use AI to its best of its ability. But don't replace those relationships that you have, in particular, in the CX side, because I think that's going to become a problem. People are going to continually gravitate to people that they have working relationships with. And AI can't replace that. I mean, I think that's great business advice across the board, not just in CX, right?

Like, I nurture, business has always been about relationships and nurturing those authentically as yourself and not as a SMS AI bot that just automatically sends things out to people even talk to recently, you know? It's just like the company Christmas card that gets sent out that you know that your CEO didn't actually look at or approve. You know, it doesn't mean that much. It reminds me a Christmas vacation and the holiday. I just watched it, which is why I said that.

Exactly. Exactly. Man, I'm in the Christmas food, apparently. I had one other question for you. This is something that's been kind of in debate, you know, on LinkedIn. They like to have strong debates about this. Or I've heard a couple different sides of this on a show before around expertise, right? So we talked about, you said data science, for example.

Like that's something that's, you know, going to still be important to understand, but I may not need to know how to actually use the tools to do it, right? Right. And I went to school, weirdly enough, for mechanical engineering, love engineering, love building. And I always felt like, yeah, the tools are going to change, right? Like, the tools I was learning engineering school that they use now are not the same. But if I can understand the ideas behind it and how, you know, the physics even, like, I need to

know that stuff so I could still build with whatever I'm going to be building with, no matter how the tool changes. Now we're at the point where I could speak into an AI tool and it would just make what I want. Right. And so then you could argue, okay, does expertise even matter? Because I don't need to know the math or the physics or anything behind any of this because the AI bot can just do it. My argument is that you can't prompt that tool properly unless you have the expertise and know what to ask of it. So I'm curious where you stand on this like how important is expertise or even like years of experience in a domain, as we get into this future where like, AI can pretty much do everything. So why would I need to understand that?

Well, I mean, I think my analogy is just a pilot on an airplane. Airplanes today can land themselves. They're doing testing right now so they can do auto takeoffs. Do I want to be on a plane without a pilot that actually knows aeronautics and understand how to actually operate that plane. That's not a plane I want to be on. So the expertise, even if in the world of airlines, you know, planes can fly themselves, you want someone on there that understands it. You know, I was talking to someone the other day and we were talking about, you know, coding and the AI, you know, AI is going to replace, you know, are we going to need engineers that write code? And my point to them was, you know, AI is going to write the code. That's great. If an engineer can't

understand the logic of that code and how it's going to work. How are we ever going to really understand if we're getting to the right outcomes and the right results? So first of all, being able to tell AI, you know, here's the things that we're looking for and here's the design and here's the intended outcome and here's how I want it to work. Being able to convey that to AI so that it can generate the code is one thing. But someone's got to look at, you know, that output over and say, hey, is this getting to our intended and intended outcomes. So I don't, I don't at all think that we're in a position where we can let the expertise go on how to fly a plane or how to write code or, you know, how to build something

mechanically. If we don't, if we don't have that, we're not going to be able to direct, you know, the tools to get to the right outcome. And we're certainly not going to be able to validate that we're getting to that right outcome. So, yeah, not a world that I, not a world that I'm looking forward to. No, I also watched Idiocracy recently, and I was thinking about this very thing. It's like, how could you possibly be an engineer if you've not learned engineering and done it? So, yeah, that's fun. Well, Mike, I'm just going to pause really quickly.

We're coming up on time. I want to end on a question about trends, but I'm going to tee it up kind of weirdly because we are in the end of 2020 or 2025. This won't be releasing until 2026. So I'm going to say this year, I mean 2026. Okay, just wanted to flag that. Okay, so it's the beginning of the year. We've got, you know, 12 months ahead of us. If I were to talk to you at the end of the year, Mike,

what's one trend that you think you would be like so excited by or just one thing that you're like, this year, you know, basically your prediction for like the end of the year something that you think will have happened? So, I mean, great question. I think by the end of year, what we will see is we will see in companies and folks in companies really developing large scale number of agents to do tasks for the company, right? And so I think, you know, the tooling's there. In many cases,

the data is already there. The AI is there. And so now it's really figuring out what are the agents that are going to have the biggest return and whether that's on customer experience or growing revenue or answering support questions, whatever it is. I think what we'll see is, a year of agent building and getting outcomes and, you know, moving more and more to autonomous agents that are, that you have the confidence in and the trust in, the actual handle particular interactions. So we're going to get there. But it's, it's not going to be overnight. It's going to be a year of, of building and trying and testing. But we will get to some really good outcomes. We're, you know, we're already saw some as we, as we ended last year. And I think that moment

and I'm just going to really continue to build. Absolutely. With, you know, autonomous agents, I want to, I would like to hear how you would define them because I feel like for the past year, we've talked about them, but it's really under the hood just been automations and not an agent.

It's just connected automations. So when we say, you know, AI agents that we trust, what does that actually mean? Let me give you, let me give you an example. So we, and it's a use case that I talked about we've done internally here, is so what we've built is we've built this account planning agent for our go-to-market team. And what that agent is doing, you know, one of the things our concept is agents never sleep. Agents are always going to be running. And so our agents are running and what they're doing for all of our customers and our customer base.

I mean, they're looking at, okay, who filed anything publicly today and what can we learn from it? You know, what's the data that we've been capturing in the last 24 hours from our telemetry systems as we're monitoring our customer systems? What's new happened in our service now system or our sales force system? And that agent is just a ton and it's making a determination. Do I need to spin up a new account plan for my rep? Because there's been a significant enough of change in any one of these, you know, different areas that, hey, the account, the account plan should change because we've learned something new. And so I think that's a great example of how, you know, this thing is just in the background running autonomously. It's making decisions on, hey, we are going to build a new account plan for company XYZ because they've stated some changes in their last earnings report.

And we're seeing some different things in terms of their telemetry data. And so, you know, that's what we're looking for more and more and more. So it's, you know, the account rep comes in. Hey, I got an email. I got a new account plan. because something's out there. You know, our agent, our account plan agent has said, it's time for a new account plan

because something's changed in the course and direction of this company. Yeah, that's great. That's great. I mean, it's really cool to hear how that's already happening. And so, like, what will that be in 12 months and what will be in 24 months? We can only guess. But I'm really excited for this world.

Yeah, I'm really excited for this world of AI agents being able to operate in the background. Truly. we got to speak to Vajoy from Cisco at Outshift or Outshift by Cisco. And he was talking about the internet of agents and agents being able to communicate with each other and like hire each other. And like the future that we have for Agentic AI is is very promising and interesting. It was definitely overhyped this year. But I think next year we'll be able to get there. Yeah.

I mean, one of the ones we're working on a POC right now is how do we have an agent that's a buyer and an agent that's a seller negotiated contracts? And so how do each agent get the best terms, whether you're a buyer or seller? And how do they get an agreement on, okay, here's the deal that we should move forward with. That's such an interesting idea, too, to think about like negotiations of the future or not even going to be done by, you and me. It's just going to be our two little agents going back and forth, trying to figure out the best terms. And then we'll be emotional about it. No, no. You didn't respond to my email fast enough, so actually don't like you now.

So I don't want to work with you. Exactly. Well, Mike, this has been fantastic. If anyone's interested in learning more about Teradata, what you guys are up to, where should they go? Hey, Teradata.com is the place to go. You can learn all about Teradata. You can actually use what's called our ClearScape experience. There's lots of ways to actually get into the software, see the capabilities, learn about how you can build your own agents.

So come on out. Awesome. All right. Thanks, Mike. All right. Thank you. It's been a great day.