6. Data Quality is More Important Than Ever in an AI World with Qlik – Tech Field Day Podcast

In our AI dominated world, data quality is the key to building useful tools. This episode of the Tech Field, a podcast features Drew Clark from Q Click, discussing best practices for integrating data sources with AI models with Joey Danone, Gina Rosenthal, and Steven Foskett. That's he, Hey, that's me.

Before Q, click Connect in Orlando. Welcome to the Tech Field Day podcast, the only podcast that dares to be both on topic or on premise. And yes, sometimes on location or on premises.

Each time we meet, we bring together a group of independent technical IT experts to discuss a single idea about a key concept in the industry. On this episode, presented by Qlik, we are talking about the importance of traditional data practices, things like data hygiene, data quality in an AI dominated world. Before we get into that though, let's meet who's on the panel for today's discussion.

I'm Joey Danton, I'm principal consultant at Denny Turing Associates Consulting. I do all things data and cloud. Hi, I am Gina Rosenthal, and I'm the CEO of Digital Sunshine Solutions.

I also have a podcast called Tech Aunties, where we talk a lot about AI hype. And I'm Drew Clark. I'm the general manager here at Click looking after the foundation of data for AI and analytics.

And I'm Steven Foskett, the organizer and president of Tech Field Day event series, part of the Futurum Group. And I am thrilled to say that we are taking Tech Field Day on the road to click connect. And so we are recording this episode of the Tech Field Day podcast as a way to let you know what's gonna go on there, what we're gonna be talking about there, and sort of what Click connect is all about.

Basically this premise, this topic is what click is all about these days. We live in an AI dominated world. It's all about ai.

Everyone would say that, uh, you know, you look at the news and and it's, it's just everywhere. But the thing is, I think a lot of people are kind of becoming wary and realizing that, uh, generative ai, especially large language models, well they're not all that great at doing things, you know, like getting the right answer to things. There is a way to make them pretty great at doing things, and that's to give them the tools that they need to do those tasks.

So, for example, give an LLMA calculator to come up with correct mathematical outcomes, give it access to a database, give it access to a list of products, or a list of customer support needs or whatever it is. Well, that's really, I think, the direction that the whole AI industry is, is headed. And frankly, that's click's bread and butter.

So, drew, talk to us a little bit here about, uh, traditional data practices and data quality. Yeah, thank you Steven. And, you know, we have thousands of customers who are using our product portfolio for kind of getting data, improving the quality of that data and getting it to where it needs to go for the decision making.

And we do this with a heritage of products from Talend, from Attunity Plus. Uh, click on what we bring to the table. So the principles, you know, as we think about, you know, the completeness of data, the confidence of kind of the timeliness that comes with that data, and as you look at and understanding where it came from or what you might need to do with it, you could be finding personal identifiable information you need to clean out unit might need to, uh, complete and kind of normalize and bring that up so that you have the right information to make your decisions or even better for getting the outcome for what you need it to do.

So the principles of data quality have been tried and true, uh, but now I think they become even more important in the velocity of this generative ai. And what you're talking about with the LLM, just this velocity of use of data is just gonna go through, uh, the roof, right? So many people.

But you need that foundation still here. One of the challenges I think is, uh, as we integrate data into LLMs, or not just LLMs, SLMs small language models, uh, within enterprises. So when we take that, you know, dependence from using an open model like chat, GPT or one of the other, or one of the various other models and try to integrate it with some of our data, is those data, those data catalog elements that we might have, like that metadata that's, uh, describing our data, protect, helping protect our data, what happens to that?

And does that get incorporated into the model? Is the model aware of that? Do we retain data security information, uh, with all sorts of regulations like GDPR and, and the California Data Privacy Act?

That's a big concern for a lot of organizations as they integrate this data into their models. What happens to the data and how is it protected and what's the security? And not only beyond that, you know, ensuring that they're going over private endpoints and not just using the free version of a cloud service, which may expose their, expose their private data to the internet, which is never a good thing.

Uh, so those are all kind of my concerns in the space around how this is gonna happen. So I think that's interesting because you're concerned about it 'cause you're a data person. I'm concerned about it 'cause I'm a storage security person.

But when you look at people bringing the, um, if you look at people bringing the, uh, AI use cases in their companies to the forefront, they just hear these fabulous, awesome things that AI can do. It should make things faster. We can find it, we can ask it a question, it'll tell us what to do.

We don't need anymore customer service reps. We don't need any more marketers. It'll be great.

Um, so they're only seeing the promise of AI and what it can do to save them money or to get things done faster. So I think there's a little bit of a disconnect between what the business hopes to accomplish and the timeframe they hope to accomplish it. And then once again, the storage and data people are gonna walk in there slowing everybody down because we're like, hold up a minute, let's think this through.

Yeah. But Gina, when you, when you talk about that, that, you know, governing is a word that sometimes people think is a bad word. Uh, but governance also can allow for, you know, the velocity to increase and with the confidence.

'cause if you don't do that, or in what Joey you were just talking about, of the metadata that goes around, kind of the information becomes so important. But if it, if it is provided and wrong and then your business users get wrong answers and you lose confidence so fast, which would then start all the way back the beginning, or I'm not gonna use it again or, and so I think there's a, um, sometimes, you know, going slow is, you know, smooth and smooth is fast, not fast is fast, right? I think I was thinking of that as you were describing it, Gina.

No, I agree with you. But I'm wondering, 'cause you talked to so many customers and help so many customers about this, do you ever come into a situation where they have gone too fast and how do you help them roll back and get everything established the right way? Yeah, it's, um, well actually one of the examples I'll give is a for, uh, is a, um, so one of the examples I'd like to kind of bring up is a auto manufacturer or a customer of ours who uses BI on ai, right?

So first stop was that they were using analytics, just what are all the queries? What's the information, what's being used? And then comparing the outcomes, you know, through these, uh, it was small language models into LLMs and you know, that first step of putting on a diagnostic tool in advance of that allowed them to roll forward, roll backwards of kind of their journey on the, the information.

And that's, and that's more of just the experiment kind of versions that we're seeing, uh, many companies go through. And I bring that up as an example of how do you throttle going back and forth on, uh, on a, um, on a, on that side and roll backing. I mean that, that becomes more of like a data science kind of, there's a lot of great disciplines around looking at what data sets.

You look at model drift. You look at the principles of that as you know, and you keep checking to make sure that you're within your normal bounds of expected results in the anomalies. We can go all the way back to those principles and apply them in this new world of high velocity, meaning high demand for these kind of new use cases on generative ai.

I think sometimes we get a little ahead of the market because we see where things are going. We technologists kind of see where things are going. But I do think that there's, uh, in addition to the AI hype, I think there's a lot of AI fear that's permeating the regular world as well.

Um, I actually had a, a technical but not our kind of technical person ask me, what is a hallucination and why do they talk about that with with ai? And I was like, I wanted to give 'em a hug because I was like, yes, yes. You're understanding.

It's not just hype and it's not just awesome and it's not just throw AI at it and everything will work. There's things we need to worry about. And I think that what you're kind of pointing at there, Gina, and, and, and Drew is that as people roll this out, they're gonna learn probably the hard way that, you know, you can't just throw a language model at it and say, we're good.

You have to give it the tools and the data, right? I think one of the things we see in any tech hype cycle, and I'm gonna go back to the big data hype cycle of the early 2010s, is we see a lot of organizations, especially smaller and mid-size organizations that wanna adopt the latest new tech, but they don't necessarily have the foundational tech that they need in place to support that. And the good example I like to use, 'cause this ties into big data and to an extent I think it ties into a lot of the, the business intelligence use cases that a lot of clicks customers use and, and frankly a lot of, uh, folks are looking for in their LLMs and that's having a solid data dimensional model, uh, where they really understand their business data and they can easily slice and dice their data.

We work with a number of organizations and yeah, I'm always amazed at how many organizations struggle to actually report, you know, what their sales are by product, by skew, by quarter, by sales rep because they don't have that model in place. And if you don't have that kind of data, uh, to support your your AI project, you're not gonna be able to get good answers. 'cause the, the model is just not gonna have the data to support it.

I do like the concept of what Drew said, uh, to talk about, uh, building out a BI model for their AI model as a start. But I think a lot of organizations have to have good BI in place before they even start to think about ai. So Joey, you're bringing up, um, in, in talk as you were describing that on a data model.

One of the other, uh, and I agree, having the discipline and understanding and, and the confidence of the fidelity of the information, you know, to use a Westworld, uh, metaphor of checking. Is it, you know, is it a real, is a copy, the is close to the real, uh, and the source is possible is, uh, a trend in what we're seeing is moving just the, the structured data, the count or the aggregate, uh, which would be done at that data model level and say, okay, this, you know, you're a customer account balance, right? Which is an account, uh, a combination of a number of kind of sources to it.

How do you make sure that that data is in your, uh, LLM via rag environment but just moving and having that uh, data point over there so that when the question comes, it's the right answer. But well, the account balance changes to, you know, based on, you know, movement. So you're not gonna keep adding in new, what you want to do is replace, so there's a change data capture kind of flow and you're not gonna blow out the size of your LLM, but you would just say and using, there's metadata and some of the tags and techniques of, alright, uh, if I'm moving from here on a mainframe data, which might have that account balance or a trade, you know, confirmation and I've centered over to the LLM up the new one, you have a little tag that goes and then you update and you just update that one record and now you know that the confidence is there.

So it's these techniques you talk about of data warehousing and then also a change data capture in this new world of, um, is I think very interesting, uh, for what is old is new, uh, in some regards I think is what you're describing from the 2010s. Yeah, for sure. I think we've seen a large uptake probably in the last year or two.

And some of it's related to AI and some of it's not across systems, uh, for capturing change data cap using change data, capture data to feed various streams. And I think a lot of that is AI driven. Um, going back to Steven's topic, when we were at, um, south by Southwest with our podcast, that's one of the things I was very interested in is we, is how the general public is looking at ai.

And it is with this really pragmatic kind of view of it can do so many beautiful things, but, uh, we heard from lots of people whose children, um, are arts majors. Like my daughter has her MFA, so it's kind of like what happens to all the writers and, and all that. And I know this is not as technical as y'all are getting, but I think that, um, because of that we're looking at the soci, some of the societal changes are from this rushing ahead and adopting AI without going slow and doing it properly.

Well it's like, where do you put the, where do you put the human, you know, uh, thinking into it and uh, and is it, where's the value of the human being in this? Are, are we gonna be replaced? Is that what you know?

And there's a little bit of this improvement. Uh, my son is applying for jobs and he is, I have to write a cover letter and he was a bit, and I'm applying for all these different jobs and he is like, well put in the job description of where you're applying, put your resume and ask for a cover letter. You know, and it came out and I was like, okay, is that you?

And he is like, well, it's not quite me. I was like, well now, but edit it, make it you. Um, and I thought it was an interesting to, and he's, um, you know, looking at it through his eyes given, you know, I'm a nerd and dealing with data all the time.

Yeah, I think that's it. But I think that there's also, even if we look at it from the nerd part, like we were just talking about is how do you know the, the stuff we're talking about is very traditional, um, data storage, just the data hygiene, which is, you know, y'all are talking about some of the things you actually do to retrieve the data and to make it usable. But this is the boring stuff that the business doesn't wanna care about.

They want you, they wanna just run their whizzbang things that, you know, they may not have even thought about. Doesn't matter if they're using an, uh, online, um, free, I feel like this is one of the first times we've had a hype cycle that's had a real customer product that they could use immediately and get some results. So like a lot of the stuff we, like, for example, the whole big data thing, that was an IT hype cycle, but, and we were doing different things and we could present different data results and it facilitated a lot of data science and, and all of those things, but there wasn't like one thing we could deliver to the customer, to the end customer from that.

Whereas this is something that end customer can use themselves and see some results. So we're seeing that demand come more from the end customers than necessarily the IT org pushing it into the business. So it's more of a top down thing than an an IT thing, which is, I think presents some unique challenges.

'cause it's gotta, you know, let the air out of the balloon slowly, uh, to explain that, you know, this isn't exactly what we want. I mean, there was some airline in Canada who had an AI chat bot, uh, and it, it, it said their bereavement fares are like 25 or 25% of full retail. Uh, and they weren't like at all.

But because the AI bot said that according kind of the in in made them, uh, honor the price. So, uh, that's a pretty simple example of, of the way things can go wrong. But if the business adopts this without, you know, talking to the data people and the data storage people and security, there's a whole lot of risk involved.

Yeah. Yeah. It, it is interesting Joey, that point because unlike, yeah, unlike most tools, this tool looks so magical and, and is so usable right off the, you know, right off the bat that I think people are just running with it.

Um, but I think that for the most part, they're running with scissors because, um, you know, it, it's very easy to do something like what you're describing there to basically roll out, you know, fire all your support staff and, and roll out an AI assistant and say, go and then find out the issues that happen when basically, I mean, the metaphor, and it's not even a stretch of a metaphor, it may even be a little bit of the reality, is that essentially you're bringing an intelligent but completely untrained person in and telling them to do a job. And this person, in this case in AI is, is fully confident in, it's like Dunning Kruger, right? They're fully confident in their abilities, they're gonna do the job you told them to do regardless of anything.

And so the, the, my point is, so we've got this amazing tool, we're putting it all over the place and we're gonna say, go, and it's gonna fail dramatically and hilariously all over the place and hopefully not harmfully in, in everywhere. And then people are gonna say, wait a second, you mean we need to give the tool a tool? And I think that's what's happening now.

We're gonna give the tool a tool to actually do the job. Well, you're, yes, you're doing, uh, and, and to be able to have the relevant kind of information. But I was going on the, you know, thinking about how long does it take to train a human into a job?

And you go from the very specialized jobs of a medical doctor and the, you know, the years of experience. And then, uh, you know, in in your, uh, your example of, alright, all these new bots or LLM new kind of business opportunities, how do you kinda understand and curate, manage, uh, train them and evolve and then wind it back, right? And or do you have to retrain it?

Uh, I was interested maybe Gina, in your perspective of particularly in the storage, maybe side of the house of the, there's the, you know, the amount of costs that we're talking about potentially in the compute, the infrastructure, the storage, uh, and the proliferation of information. If we look at information creation, it's just gonna get faster and, and more broad. It's like that's another unintended consequence of, oh, I just, lemme throw it in, let throw it in it, Joey, the big data or her dump just dump all the data and then all of a sudden the cost starts to go, well what am I doing?

Well that's a benefit, is what am I doing with it? But the cost is storage, the usage, the compute, uh, what's your perspective, Gina? Yeah, Exactly.

The, exactly what you're saying, it's the compute part is I think worse than the storage part. And I think one of the big issues is the environmental concerns. And if anybody has, uh, initiatives, green initiatives, this will blow it out of the water.

You know, so how can you, how can you have a green initiative when you, if you're, you're contributing, like, it, it is out of control. How, how much energy and how much water the data centers have to, to use. And nobody ever talks about that, about what we're doing to our environment to, to have this, um, capability.

So, um, that, that's definitely an unintended consequence. And, um, something, you know, again, that you should look at. There's the cost of it as well, like you said.

So, you know, in all we're talking out, I think all the examples we've given have just been, you know, generative ai. So what if you're using a different type of AI and you're actually trying to do something on the edge and you're trying to do something more intense, um, what kind of data do you collect? And that goes back to, that goes back to, um, that's pretty fundamental.

What do I collect? Do I collect all the logs? Do I have the capacity to store them?

You know, and now it's, do I collect all this data and put it some case in case I can use it? Do I become a data hoarder because now I think I can build a tool to make information out of the data? Or do you say, well we can catch that information another time?

Like, so it's, it's a lot of, um, it's a lot of academic, you know, experience. Um, I say data hygiene all the time 'cause it's just the stuff you have to do. You have to, so how do you back the stuff up?

It's like, how do I protect the data in case a tornado rolls through, you know, my data center? What, what do I, what are the things I have to think about to keep the data and make it available? And do I use, what sort of services do I use?

Do I rely on a cloud service to do some of this for me? And if I do that, what kind of, how locked in will I get to that? Will I, so it's all cheap now 'cause they're getting us to adopt it.

What happens if it's not? So there's a lot of, um, there's a lot of things people need to think about as far as actually running these tools and storing the data that we want to use for these tools. Yeah, that, that's, um, you know, back Joey, I was talking about the, um, bi or business intelligence on the ai.

One of the things that we're seeing with some of our customers, uh, is, you know, putting an economic view of it. What's the compute cost for all the queries that are going in? What's being sourced where, and you see, you know, the lake house, and I'm, I'd be curious if you're seeing this, but the rise of the lake house, which is just put what data, what you need where, or even leaving it, you know, on the mainframe or SAP or the IOT device, knowing where it is, bringing it and putting a, a lens of the cost for query compute storage for all the outcomes holistically across the data state is something.

Uh, and that's one of the, the interesting things of being a data integration company like Qlik and an analytics company, is that we can put both, we have a view of what's being the pipes of what's going through and we have the analytics that can go and look down or are you putting the right things through the right pipes from an economic basis, you know? And, um, and so we're getting some very interesting kind of, uh, feedback from our customers on that. So I think one of the things, uh, there, the, the data lake model has made storage cheaper.

So customers are more inclined, uh, to, to save more data. But at the same time, uh, vectorization the process that, uh, this data gets, like ai, a ball, I guess you'd call it, uh, is a huge storage consumer. And that storage has to be pretty low latency.

'cause that kind of moves out of the, the traditional ACAST model where you're separating data from compute, you're kind of moving that vectorized data back into either relational databases or some kind of new SQL database to better support those AI queries. And that is really expensive and that does consume a lot of storage because it, it basically blows out the size of the data, uh, 20 fold. Uh, and you lose the abil in, in a lot of cases, you lose the ability to compress that data 'cause it's, it's not very duplicated.

Uh, so the storage aspect's another kind of another kind of interesting cost that we, I've seen just on the infrastructure side of this, and I think because a lot of folks are consuming data from, you know, or consuming AI models just directly through a cloud provider like OpenAI, they're not necessarily exposed to those costs, but when they get into the more advanced, uh, deployment where they're hosting their own models or running their own model on a public cloud, they're seeing some of those infrastructure costs. I, Microsoft did their earnings report last week, and I, I just heard a briefing on it today and I, I can't remember the number they're spending on data centers, but I wanna say it's like 77 billion in the first quarter. And they're, they're actually, uh, connecting all of their worldwide data centers within InfiniBand because the network, they're trying to effectively build one global, uh, cluster for, for ai and the amount of money that they're spending on that networking gear, I just can't even fathom.

Well, and the reason that they have to build that Joey is because they're trying to train ever bigger models. You know, they need more and more and more and more parameters so that the models can be trained with that. But that, it seems to me like that's a completely counter intuitive or counterproductive maybe approach to ai.

Essentially what you're saying is, I need someone who is an expert on virtually every field, every topic, every bit of information in every field, instead of, I think the, the more appropriate approach as, as Gina and Drew were just talking about, which would be, I need someone who is proficient at interaction and can look up the data in the right spot. You know, basically, I, I feel like that, you know, whether it's RAG or whether it's transfer learning or tuned models, I think ultimately that may be the, the better approach to ai. So instead of, you know, GPT five with a trillion parameters of training, maybe we need a LAMA and a database with good data hygiene and just set it loose on that.

And I think that that would solve some of the environmental problems too, because we've already seen that you can run something like a llama three on a laptop or a phone instead of having it have to run on a huge pile of GPUs. And that makes everything easier. But I wonder, hey Joey, does it have to do with the grid too, the Microsoft grid and being able to train the model on the grid or, because I think there's, I I'm not good at, I'm not up to speed on the grid plus ai, but I feel like that's maybe something too.

No, that's kind of dedicated compute for AI services. All, all, all this stuff they're building right now is just dedicated to ai. Uh, I went to a public reino talk in January on this and yeah, yeah, it's all all about ai.

Uh, so Steven, what you were talking about going into those, um, product or thinking of data as a product as opposed to a product, you know, so a product lives and breathes of, uh, alright, as I'm building out the code and I wanna, or I want to kind create this thing that somebody will buy, and then you want to have updates, you want to have models, uh, you, oh, I have new information. This rise of data as a product, you know, is something we see, uh, a lot of enterprises are doing it on their own, you know, okay, this is my customer data and there's somebody who is a product owner for that data concept, which they can bring in those micro models of being able to interrogate, but they're curating, managing, um, you know, doing DevOps for it. Uh, it's not just a feed, but this rise of a data product, um, and concept is something we're seeing.

I mean, are you, uh, the rest of the panel, are you seeing this as well? This Steve brought, I I thought you did a better job than I'm doing of articulating it, but, um, are you seeing that product owners of a dataset or a data concept? I feel like it's ebbed and fluid over the last decade, but it's definitely been ebbing or flowing more than ebbing in the last couple of years with the, with AI becoming so prevalent.

Yeah, I don't know if I've seen that. Definitely the product owners being more of service owners. So it makes sense that the service owners might be the data owners for the data that's required for their service.

Yeah, and ultimately this, it just reminds me of everything before ai, right? It's, it's, it, we were already talking about data warehouses and data lake houses, and we were already talking about good data quality and business intelligence and analytics and all these things that existed prior to the explosion of LLMs. You know, this is what companies have been doing.

You know, I don't wanna sound like a grumpy old guy, get off my lawn here, but I'd say that good data practices, uh, good data hygiene, um, it, it, it's, it's as just as relevant in the AI space as it was before. Um, and I guess in a way that's the premise of this episode, right? That, that the things we've been doing before make AI work better too.

Is that, uh, an an agreement from everybody. Yeah, for sure. With me, I, I believe you know, it's a foundation if you don't have your foundation or, or what it's gonna reward is if you have your foundation like you're describing and the principles, and then you'll be able to capture and increase the, and leverage AI faster, better, uh, than if you don't, you then there'll be a certain point of collapse, uh, that you might have to go back to fix that foundation.

So I, I agree with you, Steven. If you can't tell me what your sales data was for last quarter by by SKU and by sales rep chat, g PT can either, And AI is just, it's the next, it's the next wave of computer science. We're not doing magical mystery killer here.

This is a real thing that takes really smart people to have it, to do it. And, and we, you know, we just should realize this is the digital transformation. We talked about it for so long, we're freaking in the middle of it right now.

So everything we've been talking about, um, again, it really does come down to the same kind of discussions that, that people have been having in the data space for, for many years except now that there's this new, um, use case for all of this data, which is generative ai. Um, that's, I'm sure what we're gonna be seeing when we're attending Q click Connect. Uh, all of us are gonna be there.

Uh, what are we gonna be, what are we gonna see? What are we looking forward to, I guess Drew, uh, what what would you like to call out there that we should, uh, pay attention to? Thank you.

Uh, and come to Orlando where you're going to join, you know, thousands of your peers and data practitioners. If you're using the Talend portfolio, the click portfolio, attunity portfolio of products around data, foundation, data quality, you're gonna get your hands onto kinda what we're doing, where you're going, meet peers, connect in, and this whole theme that we've talked about, uh, we're gonna have thousands of people, uh, having the same theme. But, you know, becoming and learning how to use the tools to have the seat at the table for the next generation of ai.

I know I was looking at the agenda and there's some, there's a lot of things that look interesting, but there's two I was very interested in at the AI council. I can't wait to hear from them. I think that'll be really good.

But there's something at night like a, not a hackathon, but something, and there's pizza involved and that looked very interesting to me. So what's that all about? Yeah, that's like a hands-on.

Oh, I love the hackathons. Those are ha hands-on where we will, uh, take some data that's needed by an NGO or non-governmental organization. Uh, and in years past, it would be a refugee crisis and there's information that you know, and supply and demand, but how do you get that to the right place and you build and get your hands on the product and you're building out.

And then at the end of the night, after some beer and pizza, uh, is you make you give the prizes out, uh, for who needs it. Uh, we did a climate change one and the winning team actually went to the city of Austin and presented their the results from the hackathon. It was pretty cool.

Any big AI announcements we're expecting at Click Connect? Yeah. Uh, so we are talking about, you know, the follow through of, uh, what we've been communicating around stage our platform, uh, which is about the data foundation to the AI in the products for productivity enhancement.

So this is some, a theme that we talked about to just make the data engineers, the business analysts more productive and effective. And there'll be new, uh, AI capabilities linking, structured and unstructured data together in a lot of the ways that you've talked about as the value and the benefit, uh, but using, you know, modern, uh, capabilities, uh, combined with the heritage of what made us unique and different, uh, from our own basis. So we have some new AI capabilities that, uh, everybody will have access to, uh, that you'll see and learn more about in June.

Excellent. Well, I can't wait to be there. It's gonna be great to have the Tech Field Day delegates joining us at click, uh, click connect.

Um, I'll be there, uh, Gina and Joey will be there. And, uh, we'll be recording some Tech Field day content and sharing it with the audience, uh, from Click Connect as well. So I hope that, uh, our listeners will join us too.

Uh, thank you so much for joining us for this episode of the Tech Field Day podcast. Before we go, uh, let me just go around and, uh, let know, uh, where can we connect with you? Where can we continue this conversation?

com. com. And, uh, LinkedIn, uh, is best ways for me, uh, to connect in with everyone, and it's been great, uh, getting to know everybody here and look forward to seeing you in Orlando.

Excellent. And, uh, as for me, of course, you'll catch me here on the Tech Field Day podcast, as well as our utilizing tech podcasts on Monday, which is focused on AI data infrastructure. That sounds familiar, uh, coming soon as well as of course, our Wednesday Galt News rundown.

Thank you very much for listening to this episode of the Tech Field Day podcast. If you enjoyed it, uh, please do give us a subscription. You'll can find us on YouTube or your favorite podcast application, and that way you won't miss an episode.

Also, we'd love a rating and a review and maybe some feedback. This podcast was brought to you by Qlik as part of, uh, the runup to click Connect, as well as by Tech Field Day Home for IT experts across the enterprise. Now part of the Future Group.

com/podcast. Thanks for listening, and we will see you next week.