Qlik Connect 2024 | Drew Clarke | Qlik

theCUBE.net

Clips
News
More from Qlik Connect 2024

Drew Clarke

EVP, Product & Technology

Qlik

Harnessing unstructured data: QlikTech’s vision for AI-driven business insights

The landscape of data management and analytics is undergoing a significant transformation, driven by advancements in generative AI in data management and the effective utilization of unstructured data.As businesses strive to harness the full potential of their data, the focus has shifted to creating robust data foundations and integrating new AI-driven applications. This evolution is not only enhancing productivity, but also reshaping how organizations perceive and utilize data, according to Drew Clarke (pictured), general manager and executive vice president of the Data Business Unit at QlikTech International AB.Qlik’s Drew Clarke talks to theCUBE about generative AI in data management

play_circle_outline Highlights of the keynote and new features like Talend Cloud and productivity enhancements

play_circle_outline Adoption curve and experimenting with generative AI

play_circle_outline Overcoming Data Availability and Low Latency Challenges with Qlik Replicate and Change Data Capture

Info

Drew Clarke | Qlik

Drew Clarke

EVP, Product & Technology Qlik

search

>> Welcome back everyone to theCUBE's coverage here in Orlando, Florida. I'm John Furrier, host of theCube with Shelley Kramer. We're here at Qlik Connect 2024. I've got a great guest coming on, Drew Clarke, GM and Executive Vice President of the Data Business Unit. Drew, great to have you on. theCUBE.

Drew Clarke

>> John, glad to be here. And Shelley, nice to meet you as well.

>> .

>> Great to see you on stage. And I got to say, I'm very impressed. Obviously data AI without data is no AI. Everyone knows that now. But data, if you've done the work, generative AI is an opportunity immediately. You guys showed some demos that weren't smoking mirrors, real demos.

Drew Clarke

>> That's right.

>> Products are going to ship in less than a month from on-stage performance. You guys are in a good position and data is now the value proposition. It's not an ingredient, it is the product. The software with data is the end game. And with generative AI, a whole new category is right around the corner of new applications. You guys are in the middle of it. What's the sentiment?

Drew Clarke

>> We're so excited to be here to be able to showcase all the work we've been doing, and that work goes all the way back on decades. The same infrastructure and discipline you have to do about feeding data into an application or analytics, it's that same foundation for the new use cases like you're describing as generative AI. All right, that's a new consumption vehicle for needing that data pipeline. So we've been doing it, working on it and we're excited just to bring it all together.

>> A couple key points in the keynote you guys are highlighting here. One is obviously answers, which is the result of having all that data. We were joking on our intro keynote analysis, you go back a decade, 14 years ago or so, Hadoop was the craze. Big data, "Ask good questions and maybe you'll get answers." Now the questions aren't even in there. The data is already baked in. The answers are being generated from great data management, great data capabilities. And now you've got the Talend Cloud. So you've got the Talend Cloud and you've got now a new user experience on productivity. Talk about those two things, the Talend Cloud and how that productivity piece from the data emerges.

Drew Clarke

>> So think of it as a data foundation, and that's what Qlik Talend Cloud is. It's about getting the data from where it happens to be. When you go buy something at a department store, that transaction is written right into an SAP or a mainframe system. That foundation says, "Okay, I have to get that data to that new Qlik answer on that piece. So the way you think about it, Qlik Talend Cloud foundation for data and it's what's needed.

Drew Clarke

>> I was looking at the report that you did, you just released in conjunction with ETR who happens to be a research partner of ours as well. And when we're talking about data, one of the things that's so interesting to me, you've been in the data business for a long time, so all of the importance of data, this isn't news to you. You've been building this for 30 years. But some of the data points from the research study spoke to some real challenges here. 64% of the survey respondents said that their people never use unstructured data to derive insights. 64%, that's a lot. 70% said they barely begun to even realize the value of their data. So I think that that really is to me, what makes this such an exciting moment because you just have greenfield ahead. We were talking about TAM earlier-

>> Yeah, there's a huge TAM....

>> and you were quizzing about expected TAM, but I think this really has to be an exciting time for Qlik.

Drew Clarke

>> Well, and you're honing in on an important part of that study, which was about that unstructured data. And actually I had the opposite reaction of it was like, "Wow, that's actually a bigger number than I thought that were using it." Because when you think about unstructured data, and a lot of people, "Oh, it's X or Twitter or it's emails." But when you think about all the documents and everything that's generated by a company, it dwarfs structured data. And to be able to harness and access that and deliver it, it's a big tailwind for us is the way we think of it from needing our platforms. And more importantly, it's inspiring for our customers to see what they're doing now and how they can use our platform to accomplish their missions. And that's where I get excited.

Drew Clarke

>> Yeah.

>> And I think the unstructured data also speaks to the general AI opportunity that you guys have because now you have multimodal assets, you have unstructured, structured data, but also you've got a lot of new loose information hanging around, whether it's from IOT devices, anything that's being measured on the network can be measured now. So we talked about the sports segment earlier, analytics for player performance. You guys have a lot of customers in sports analytics. Okay, well business is now situationally aware of what's going on. Execution in sports and business are now converging. The nexus of performance is going to be the tech athlete, the business person. That's where this is going. I mean this is where the 360-

Drew Clarke

>> Yeah, it is interesting John, you're using a sports analogy because-

>> Always.

Drew Clarke

>> Oh, he's always the sport guy?

>> Always.

Drew Clarke

>> All right.

>> We had a hockey guy on earlier, we loved him.

Drew Clarke

>> Okay. Well we got it. Well, but there are also rules and implications if you don't use the data correctly. And so yes, you have access to it, but you look at in the European Union, they have the AI Act. And if you're using AI for decision-making and you can't back it up, or worse, you're using data you shouldn't be using, that's a foul. And you could get disqualified, or even worse, penalized at pretty significant ways. So there is a responsibility to using this data right. Just because you can doesn't mean you should.

>> Yeah. And you guys had a great trust message, I thought on stage was you hit that point home. And I got to call out and say, I've never seen a company on stage say on a keynote, "Data supply chain in context to the overall value proposition." That is really the unpacked supply chain. You're talking about the concept of software supply chain in a security context. So LLMs with hallucinations point to gaps in data. If data's not available, you start to see these hallucinations. And as unstructured becomes vectorized, and you guys talked a lot about vectors as well, you have now retrieval search, retrieval aspects of the data. So now everything is available. So now you have, "Okay, what does that mean for the user?" It means more data scientists are basically going to be born in the wild now with data. So it's a fundamental game changer. So how do you guys see that expanding beyond the software? Because now data is the product. Data is the product.

Drew Clarke

>> Yeah, yeah. And you need to apply product design principles. And it's almost easier if we take a step out of the software or the tech industry and go, "Okay, well let's think about consumer-oriented products." And a consumer-oriented product that you could buy at the supermarket, Crest or Procter & Gamble, and the decisions around what's in that product, how it's built, who's buying it, how do you have that response and the iteration, where did it come from? And you go all the way back and how did it get to that marketplace? Those same principles is what we can do and are doing with that data supply chain. It's not just getting it to the shelf, but when it's on the shelf, is it high up? Is it low down? How much is it being picked up? Is it being used? And the data product owners, these are new and these are new jobs. Some of them I've met here and it's like, "Wow, you're coming at it like a consumer product person." And that's the differentiation and-

>> And the trust piece, and then the explainable AI was a huge wave right now, and people don't even know what that means. But the answer is you got to be able to explain, is this data coming from authentic sources? And with synthetic data, there's a lot of new data sources out. The game is changing. The datasets are changing, the data changes. How do you guys keep up with it? Scope the problems and opportunity there for us. As you look at the landscape, more data's coming in unstructured.

Drew Clarke

>> That's right.

>> Different types of data's coming in. The rise of synthetic data is now huge. Open data formats are dominating.

Drew Clarke

>> Well, John, I would maybe think of it in a way and almost put yourself as a user and an experience, because you brought up the word trust. And you're using hallucinations, which are like, "Okay, that's generative AI spinning out and giving out random answers." But if you were to say, "All right, I'm interacting with a company and I want to know the latest information about my account balance." If the number's wrong, you lose trust pretty quickly. Like, "Wait a minute, that's not right." And that's where we're spending the time of ... And it's not just unstructured data assembling into that generative AI, it's that structured data, an account balance, statuses of a shipment. And that needs to be the most current information because if we lose your trust, why would you come back? And we realize our customers are going through that experience to say, "We have to build this new product, but what we put in there, it's got to be right."

>> Yeah, there's no tolerance. There's zero tolerance for not being accurate 100% in the enterprise. The data's being relied on day in and day out. It's not just a search query like we see in OpenAI and ChatGPTs of the world. I can dissect it, but in the enterprise you're dealing with really mission-critical workloads end to end. Another big theme.

Drew Clarke

>> That's right. Yeah.

Drew Clarke

>> I have a question. One of the things that you covered in this research study was people who indicated they don't have any plans to start using gen AI and these solutions for managing their data. And not surprisingly, 53% said, "We just don't have the expertise needed." And another 47%, "We haven't seen any value from gen AI tools." So it seems to me there's a lot of opportunity there for customer education and for ... I know that a lot of what you're focused on also is bringing training and the importance of regular training to employees, creating a culture of data and understanding what that is and all that sort of thing. So talk to us a little bit about that and what you see ahead.

Drew Clarke

>> Yeah, so Shelley, we've talked about data literacy as a company before, which is about reading, writing, understanding, and arguing with data, because literacy is more than just that. And so as we invested and we had the data literacy project, which was around that first fundamentals, well generative AI, when you talk about these use cases and the plans, are you going to use that? What people are now taking a step back on is what is the art of the possible for new ways of working with and delivering products to a customer or solution? And I think that's the part where people can see a couple of easy things of, "Well, I can improve productivity. So I can compress, shorten an amount the time to read and write, maybe understand." So okay, productivity is done. Arguing with data, that's the thinking part. That one is where people are going, "All right, well is this the right use case that's going to really help deliver a new way of delivering on our product?" So what we're doing is we're A, expanding on that data literacy project, we're B, working with our executive advisory council, the AI council you're talking to later. And we're evolving that capabilities we have where look, we have the art of the possible technologically. What is it that you're trying to do from a literacy consumption usage of that information? So that's where we're going.

Drew Clarke

>> Makes perfect sense.

>> So you're on the front lines with customers and they're looking at AI and the generative AI specifically as new category. Okay, we talked about that. They're lining up their ducks, so to speak, and try to get in that world. A little budget shift from other cloud-native or other activities into this generative AI. But you're starting to see now clear lines of sight around formation of budgets and investments. It's still light, and you guys did a survey that came out. I saw the press release. Still lower on the adoption, on the investment side.

Drew Clarke

>> Yeah, that's right.

>> So investment and adoption's coming quickly. We're seeing everyone focus on this bottoms up and top down throughout the organizations. What does the adoption curve look like? Is there indifference to generative AI? From a budget standpoint are people leaning in hard? Are they taking baby steps, to your point about accuracy? You got to get a lot of prerequisites right to get it right, and you don't want to take on any technical debt or business debt going into that. What is your view on the adoption curve and the adoption cycle?

Drew Clarke

>> John, you're exactly right. The customers I talk to, everybody is talking about generative AI or AI. And I make a point of like, "Okay, so what are you doing about it?" "Oh yeah, we have a top-down initiative. We have to do something about AI." "Okay, great. What is it?" "Well, we don't know yet."

>> Retrieval-

>> What are you trying to solve for?

>> Sectorize it.

Drew Clarke

>> And so bottoms up, what we see is a lot of experimenting going on. And so similar to other data science or machine learning cycles, there's experiments that are going on. So you've got budgets that have been put aside. Right now experiments, and then within the experiments they're looking at what's the real cost drivers? And this is why if you're going to run your own private generative AI, well, okay, so now you have to have big heavy machines, vector databases. Where do you do it? There's, "All right. Well if I do it in a public cloud, am I going to put all my information up there?" And there's different cost structures that people are adjusting and looking at. It is squarely in the experiment phase. I think we'll see is next year more defined budgets as projects come in. I do know just every company is experimenting right now. And where Qlik Talend Cloud comes in is once you experiment, then the next question is, "Do you have the trusted right data going into it?" And that's where our data supply chain fits right into that.

Drew Clarke

>> Yeah.

>> Yeah, and you guys are taking all the nice cloud-native concepts, observability, software supply chain, now data supply chain, you're starting to see the value of the data. Data marketplaces saw the announcement with AWS. Again, more partnerships there, ecosystem building. The question I want to ask you is that as the data becomes the key part of the product, is there a vision where you see Qlik's market changing from just analytics to just all applications? Because if you connect the dots, this is where it's going. It's like the data is in everything. And so if that's there, there's no category for that. That's just basically-

Drew Clarke

>> John, have you been in my office looking at my whiteboard?

Drew Clarke

>> He has actually.

>> Did I hit a nerve there?

>> He snuck in there .

Drew Clarke

>> Right. Well, and this comes back to the enterprise layer. So we're organized, as I've been outlining, as a data business unit, which we're about the foundation of data. Brendan, with the applications and the analytics and the use of that data for business outcomes. And what's converging very quickly at our largest enterprise customers is you need both. And you think about the pervasiveness of that data. And an application is not a dashboard, an application is a phone and it's in the hands of the store manager, and they're getting an alert that says, "Well, your staffing is going to blow out of your budget for overtime. Maybe you need to think about who you've got for your coverage for this coming weekend." That was a proactive data alert that went out to the store manager where they work on their phone. That's an application, that's data. It's not the traditional view. Our TAM is much broader on just exactly that.

>> Shelley's on theCUBEResearch team, and one of the things we've observed and written about and published is that in the past year at these events, the theme keeps popping back up. And if you look at NVIDIA and the silicon work being done around GPUs and NPUs, XPUs in general on the hardware side, all the performance is targeting these end-to-end workloads. And so what we're seeing is the concept of end-to-end workloads with data as the new intellectual property. And companies are, I hate to bring back the old the data is an asset on the balance sheet, but you're starting to see that workflows that are scoped and well understood and the data behind them are going to be retrofitted and reset with generative AI in mind, hence changing the game on what data's in there. So if you believe that to be true, then you say, "Okay, that's an intellectual property of the company." So then everything optimizes around that. What's your reaction to that? Do you believe that to be the case? And if so, how are companies looking at that and preparing themselves as they build the foundation to lock in capabilities to support that?

Drew Clarke

>> Okay, so there are two things there, John, as you're describing that I was thinking through. One is as an enterprise and you think about the architectures that they have to build up, and the medallion architecture's pretty common now for lake houses. And that's like a bronze tier, which is your raw flat data, your silver, your gold, and classically it's a data lake, a warehouse, to a data mart. All right, that needs to be reimagined, as you think about at the top of that generative AI is like, "Well, what goes where, and you have it in that pipeline?" But you brought up NVIDIA and the chips and the GPUs, because if you just take that whole area, you take it up a level, and underneath is the compute and the compute infrastructure that's needed to power all of that. And you're going to see this verticalization, I think in the next ... and this is my opinion. Over the next 10 years, of verticalization just going from silica, and then who knows how quantum is going to really change the game on this through the infrastructure to the velocity of what's used where. And there's a new announcement from Elon's company with a chip in the brain they're going to use.

Drew Clarke

>> Yeah, Neuralink. Don't you have one?

>> He's got all the answers. He must have one.

Drew Clarke

>> . But that verticalization from silica all the way to outcome, I don't think it's as far away as we think it is.

>> No.

Drew Clarke

>> It's going to be in our lifetimes for sure.

>> It's dizzyingly how quickly all of this is evolving. It's so exciting. And every time I say that, somebody says, "Yes, it is exciting," but terrifying also works.

Drew Clarke

>> Yeah, that's right.

>> But Drew, great to have you on. We can go another hour and go into the weeds, but one last question on the real-time. I noticed that's a big part of the announcement. Data's only as good if it gets to the AI in the application. So making data low latency available is going to be a big challenge. How do you see that evolving very quickly? What's your view vision on data availability or highly available data?

Drew Clarke

>> That word, that phrase, right?

>> Which word do we use? Highly available or high availability? This is a storage data availability problem.

Drew Clarke

>> Well, so this is actually where the ... And I'm thrilled to be representing and part of the data team that has change data capture, the background that we have. And this goes back to Attunity and there's 20 years of working at. As data has changed, or that order I told you, a little data point is done there, and then that has to go all the way through the chain and get to this point. You just need to move what changed, so change data capture is done. What Attunity has and with Qlik Replicate is a low latency means we're looking at the log files of what's changing. We don't have to bring down that source system, and then we just move what's changed over to where it needs to. And earlier today on live broadcast, Vanguard talked about how much data they're moving in real-time to where it needs to. That's Qlik Replicate. That's our background. We have this running at scale. So this isn't new. This is actually, we've been doing this for a long time with scale.

>> Good job. Final question to end the segment. How excited are you around the fact that generative AI's here, and how would you explain to someone watching what's different about generative AI with Qlik going forward? What's going to be the big change in value and user experience? How would you describe what's different in this new wave of generative AI?

Drew Clarke

>> What's different is that individuals will be able to do their jobs faster, better using the trusted data that comes from a great foundation, but is combined unstructured and structured. You get the right account balance, but they also know, and then can have a conversation with you about that balance. That's what's different.

Drew Clarke

>> Yeah.

>> Drew Clarke, General Manager, Executive Vice President here on theCUBE with the data business unit at Qlik. We're bringing you all the data here live on theCUBE. Stay with us for more coverage from Orlando from Qlik Connect 24. We'll be back after this short break.

Drew Clarke | Qlik

search

>> Welcome back everyone to theCUBE's coverage here in Orlando, Florida. I'm John Furrier, host of theCube with Shelley Kramer. We're here at Qlik Connect 2024. I've got a great guest coming on, Drew Clarke, GM and Executive Vice President of the Data Business Unit. Drew, great to have you on. theCUBE.

Drew Clarke

>> John, glad to be here. And Shelley, nice to meet you as well.

>> .

>> Great to see you on stage. And I got to say, I'm very impressed. Obviously data AI without data is no AI. Everyone knows that now. But data, if you've done the work, generative AI is an opportunity immediately. You guys showed some demos that weren't smoking mirrors, real demos.

Drew Clarke

>> That's right.

>> Products are going to ship in less than a month from on-stage performance. You guys are in a good position and data is now the value proposition. It's not an ingredient, it is the product. The software with data is the end game. And with generative AI, a whole new category is right around the corner of new applications. You guys are in the middle of it. What's the sentiment?

Drew Clarke

>> We're so excited to be here to be able to showcase all the work we've been doing, and that work goes all the way back on decades. The same infrastructure and discipline you have to do about feeding data into an application or analytics, it's that same foundation for the new use cases like you're describing as generative AI. All right, that's a new consumption vehicle for needing that data pipeline. So we've been doing it, working on it and we're excited just to bring it all together.

>> A couple key points in the keynote you guys are highlighting here. One is obviously answers, which is the result of having all that data. We were joking on our intro keynote analysis, you go back a decade, 14 years ago or so, Hadoop was the craze. Big data, "Ask good questions and maybe you'll get answers." Now the questions aren't even in there. The data is already baked in. The answers are being generated from great data management, great data capabilities. And now you've got the Talend Cloud. So you've got the Talend Cloud and you've got now a new user experience on productivity. Talk about those two things, the Talend Cloud and how that productivity piece from the data emerges.

Drew Clarke

>> So think of it as a data foundation, and that's what Qlik Talend Cloud is. It's about getting the data from where it happens to be. When you go buy something at a department store, that transaction is written right into an SAP or a mainframe system. That foundation says, "Okay, I have to get that data to that new Qlik answer on that piece. So the way you think about it, Qlik Talend Cloud foundation for data and it's what's needed.

Drew Clarke

>> I was looking at the report that you did, you just released in conjunction with ETR who happens to be a research partner of ours as well. And when we're talking about data, one of the things that's so interesting to me, you've been in the data business for a long time, so all of the importance of data, this isn't news to you. You've been building this for 30 years. But some of the data points from the research study spoke to some real challenges here. 64% of the survey respondents said that their people never use unstructured data to derive insights. 64%, that's a lot. 70% said they barely begun to even realize the value of their data. So I think that that really is to me, what makes this such an exciting moment because you just have greenfield ahead. We were talking about TAM earlier-

>> Yeah, there's a huge TAM....

>> and you were quizzing about expected TAM, but I think this really has to be an exciting time for Qlik.

Drew Clarke

>> Well, and you're honing in on an important part of that study, which was about that unstructured data. And actually I had the opposite reaction of it was like, "Wow, that's actually a bigger number than I thought that were using it." Because when you think about unstructured data, and a lot of people, "Oh, it's X or Twitter or it's emails." But when you think about all the documents and everything that's generated by a company, it dwarfs structured data. And to be able to harness and access that and deliver it, it's a big tailwind for us is the way we think of it from needing our platforms. And more importantly, it's inspiring for our customers to see what they're doing now and how they can use our platform to accomplish their missions. And that's where I get excited.

Drew Clarke

>> Yeah.

>> And I think the unstructured data also speaks to the general AI opportunity that you guys have because now you have multimodal assets, you have unstructured, structured data, but also you've got a lot of new loose information hanging around, whether it's from IOT devices, anything that's being measured on the network can be measured now. So we talked about the sports segment earlier, analytics for player performance. You guys have a lot of customers in sports analytics. Okay, well business is now situationally aware of what's going on. Execution in sports and business are now converging. The nexus of performance is going to be the tech athlete, the business person. That's where this is going. I mean this is where the 360-

Drew Clarke

>> Yeah, it is interesting John, you're using a sports analogy because-

>> Always.

Drew Clarke

>> Oh, he's always the sport guy?

>> Always.

Drew Clarke

>> All right.

>> We had a hockey guy on earlier, we loved him.

Drew Clarke

>> Okay. Well we got it. Well, but there are also rules and implications if you don't use the data correctly. And so yes, you have access to it, but you look at in the European Union, they have the AI Act. And if you're using AI for decision-making and you can't back it up, or worse, you're using data you shouldn't be using, that's a foul. And you could get disqualified, or even worse, penalized at pretty significant ways. So there is a responsibility to using this data right. Just because you can doesn't mean you should.

>> Yeah. And you guys had a great trust message, I thought on stage was you hit that point home. And I got to call out and say, I've never seen a company on stage say on a keynote, "Data supply chain in context to the overall value proposition." That is really the unpacked supply chain. You're talking about the concept of software supply chain in a security context. So LLMs with hallucinations point to gaps in data. If data's not available, you start to see these hallucinations. And as unstructured becomes vectorized, and you guys talked a lot about vectors as well, you have now retrieval search, retrieval aspects of the data. So now everything is available. So now you have, "Okay, what does that mean for the user?" It means more data scientists are basically going to be born in the wild now with data. So it's a fundamental game changer. So how do you guys see that expanding beyond the software? Because now data is the product. Data is the product.

Drew Clarke

>> Yeah, yeah. And you need to apply product design principles. And it's almost easier if we take a step out of the software or the tech industry and go, "Okay, well let's think about consumer-oriented products." And a consumer-oriented product that you could buy at the supermarket, Crest or Procter & Gamble, and the decisions around what's in that product, how it's built, who's buying it, how do you have that response and the iteration, where did it come from? And you go all the way back and how did it get to that marketplace? Those same principles is what we can do and are doing with that data supply chain. It's not just getting it to the shelf, but when it's on the shelf, is it high up? Is it low down? How much is it being picked up? Is it being used? And the data product owners, these are new and these are new jobs. Some of them I've met here and it's like, "Wow, you're coming at it like a consumer product person." And that's the differentiation and-

>> And the trust piece, and then the explainable AI was a huge wave right now, and people don't even know what that means. But the answer is you got to be able to explain, is this data coming from authentic sources? And with synthetic data, there's a lot of new data sources out. The game is changing. The datasets are changing, the data changes. How do you guys keep up with it? Scope the problems and opportunity there for us. As you look at the landscape, more data's coming in unstructured.

Drew Clarke

>> That's right.

>> Different types of data's coming in. The rise of synthetic data is now huge. Open data formats are dominating.

Drew Clarke

>> Well, John, I would maybe think of it in a way and almost put yourself as a user and an experience, because you brought up the word trust. And you're using hallucinations, which are like, "Okay, that's generative AI spinning out and giving out random answers." But if you were to say, "All right, I'm interacting with a company and I want to know the latest information about my account balance." If the number's wrong, you lose trust pretty quickly. Like, "Wait a minute, that's not right." And that's where we're spending the time of ... And it's not just unstructured data assembling into that generative AI, it's that structured data, an account balance, statuses of a shipment. And that needs to be the most current information because if we lose your trust, why would you come back? And we realize our customers are going through that experience to say, "We have to build this new product, but what we put in there, it's got to be right."

>> Yeah, there's no tolerance. There's zero tolerance for not being accurate 100% in the enterprise. The data's being relied on day in and day out. It's not just a search query like we see in OpenAI and ChatGPTs of the world. I can dissect it, but in the enterprise you're dealing with really mission-critical workloads end to end. Another big theme.

Drew Clarke

>> That's right. Yeah.

Drew Clarke

>> I have a question. One of the things that you covered in this research study was people who indicated they don't have any plans to start using gen AI and these solutions for managing their data. And not surprisingly, 53% said, "We just don't have the expertise needed." And another 47%, "We haven't seen any value from gen AI tools." So it seems to me there's a lot of opportunity there for customer education and for ... I know that a lot of what you're focused on also is bringing training and the importance of regular training to employees, creating a culture of data and understanding what that is and all that sort of thing. So talk to us a little bit about that and what you see ahead.

Drew Clarke

>> Yeah, so Shelley, we've talked about data literacy as a company before, which is about reading, writing, understanding, and arguing with data, because literacy is more than just that. And so as we invested and we had the data literacy project, which was around that first fundamentals, well generative AI, when you talk about these use cases and the plans, are you going to use that? What people are now taking a step back on is what is the art of the possible for new ways of working with and delivering products to a customer or solution? And I think that's the part where people can see a couple of easy things of, "Well, I can improve productivity. So I can compress, shorten an amount the time to read and write, maybe understand." So okay, productivity is done. Arguing with data, that's the thinking part. That one is where people are going, "All right, well is this the right use case that's going to really help deliver a new way of delivering on our product?" So what we're doing is we're A, expanding on that data literacy project, we're B, working with our executive advisory council, the AI council you're talking to later. And we're evolving that capabilities we have where look, we have the art of the possible technologically. What is it that you're trying to do from a literacy consumption usage of that information? So that's where we're going.

Drew Clarke

>> Makes perfect sense.

>> So you're on the front lines with customers and they're looking at AI and the generative AI specifically as new category. Okay, we talked about that. They're lining up their ducks, so to speak, and try to get in that world. A little budget shift from other cloud-native or other activities into this generative AI. But you're starting to see now clear lines of sight around formation of budgets and investments. It's still light, and you guys did a survey that came out. I saw the press release. Still lower on the adoption, on the investment side.

Drew Clarke

>> Yeah, that's right.

>> So investment and adoption's coming quickly. We're seeing everyone focus on this bottoms up and top down throughout the organizations. What does the adoption curve look like? Is there indifference to generative AI? From a budget standpoint are people leaning in hard? Are they taking baby steps, to your point about accuracy? You got to get a lot of prerequisites right to get it right, and you don't want to take on any technical debt or business debt going into that. What is your view on the adoption curve and the adoption cycle?

Drew Clarke

>> John, you're exactly right. The customers I talk to, everybody is talking about generative AI or AI. And I make a point of like, "Okay, so what are you doing about it?" "Oh yeah, we have a top-down initiative. We have to do something about AI." "Okay, great. What is it?" "Well, we don't know yet."

>> Retrieval-

>> What are you trying to solve for?

>> Sectorize it.

Drew Clarke

>> And so bottoms up, what we see is a lot of experimenting going on. And so similar to other data science or machine learning cycles, there's experiments that are going on. So you've got budgets that have been put aside. Right now experiments, and then within the experiments they're looking at what's the real cost drivers? And this is why if you're going to run your own private generative AI, well, okay, so now you have to have big heavy machines, vector databases. Where do you do it? There's, "All right. Well if I do it in a public cloud, am I going to put all my information up there?" And there's different cost structures that people are adjusting and looking at. It is squarely in the experiment phase. I think we'll see is next year more defined budgets as projects come in. I do know just every company is experimenting right now. And where Qlik Talend Cloud comes in is once you experiment, then the next question is, "Do you have the trusted right data going into it?" And that's where our data supply chain fits right into that.

Drew Clarke

>> Yeah.

>> Yeah, and you guys are taking all the nice cloud-native concepts, observability, software supply chain, now data supply chain, you're starting to see the value of the data. Data marketplaces saw the announcement with AWS. Again, more partnerships there, ecosystem building. The question I want to ask you is that as the data becomes the key part of the product, is there a vision where you see Qlik's market changing from just analytics to just all applications? Because if you connect the dots, this is where it's going. It's like the data is in everything. And so if that's there, there's no category for that. That's just basically-

Drew Clarke

>> John, have you been in my office looking at my whiteboard?

Drew Clarke

>> He has actually.

>> Did I hit a nerve there?

>> He snuck in there .

Drew Clarke

>> Right. Well, and this comes back to the enterprise layer. So we're organized, as I've been outlining, as a data business unit, which we're about the foundation of data. Brendan, with the applications and the analytics and the use of that data for business outcomes. And what's converging very quickly at our largest enterprise customers is you need both. And you think about the pervasiveness of that data. And an application is not a dashboard, an application is a phone and it's in the hands of the store manager, and they're getting an alert that says, "Well, your staffing is going to blow out of your budget for overtime. Maybe you need to think about who you've got for your coverage for this coming weekend." That was a proactive data alert that went out to the store manager where they work on their phone. That's an application, that's data. It's not the traditional view. Our TAM is much broader on just exactly that.

>> Shelley's on theCUBEResearch team, and one of the things we've observed and written about and published is that in the past year at these events, the theme keeps popping back up. And if you look at NVIDIA and the silicon work being done around GPUs and NPUs, XPUs in general on the hardware side, all the performance is targeting these end-to-end workloads. And so what we're seeing is the concept of end-to-end workloads with data as the new intellectual property. And companies are, I hate to bring back the old the data is an asset on the balance sheet, but you're starting to see that workflows that are scoped and well understood and the data behind them are going to be retrofitted and reset with generative AI in mind, hence changing the game on what data's in there. So if you believe that to be true, then you say, "Okay, that's an intellectual property of the company." So then everything optimizes around that. What's your reaction to that? Do you believe that to be the case? And if so, how are companies looking at that and preparing themselves as they build the foundation to lock in capabilities to support that?

Drew Clarke

>> Okay, so there are two things there, John, as you're describing that I was thinking through. One is as an enterprise and you think about the architectures that they have to build up, and the medallion architecture's pretty common now for lake houses. And that's like a bronze tier, which is your raw flat data, your silver, your gold, and classically it's a data lake, a warehouse, to a data mart. All right, that needs to be reimagined, as you think about at the top of that generative AI is like, "Well, what goes where, and you have it in that pipeline?" But you brought up NVIDIA and the chips and the GPUs, because if you just take that whole area, you take it up a level, and underneath is the compute and the compute infrastructure that's needed to power all of that. And you're going to see this verticalization, I think in the next ... and this is my opinion. Over the next 10 years, of verticalization just going from silica, and then who knows how quantum is going to really change the game on this through the infrastructure to the velocity of what's used where. And there's a new announcement from Elon's company with a chip in the brain they're going to use.

Drew Clarke

>> Yeah, Neuralink. Don't you have one?

>> He's got all the answers. He must have one.

Drew Clarke

>> . But that verticalization from silica all the way to outcome, I don't think it's as far away as we think it is.

>> No.

Drew Clarke

>> It's going to be in our lifetimes for sure.

>> It's dizzyingly how quickly all of this is evolving. It's so exciting. And every time I say that, somebody says, "Yes, it is exciting," but terrifying also works.

Drew Clarke

>> Yeah, that's right.

>> But Drew, great to have you on. We can go another hour and go into the weeds, but one last question on the real-time. I noticed that's a big part of the announcement. Data's only as good if it gets to the AI in the application. So making data low latency available is going to be a big challenge. How do you see that evolving very quickly? What's your view vision on data availability or highly available data?

Drew Clarke

>> That word, that phrase, right?

>> Which word do we use? Highly available or high availability? This is a storage data availability problem.

Drew Clarke

>> Well, so this is actually where the ... And I'm thrilled to be representing and part of the data team that has change data capture, the background that we have. And this goes back to Attunity and there's 20 years of working at. As data has changed, or that order I told you, a little data point is done there, and then that has to go all the way through the chain and get to this point. You just need to move what changed, so change data capture is done. What Attunity has and with Qlik Replicate is a low latency means we're looking at the log files of what's changing. We don't have to bring down that source system, and then we just move what's changed over to where it needs to. And earlier today on live broadcast, Vanguard talked about how much data they're moving in real-time to where it needs to. That's Qlik Replicate. That's our background. We have this running at scale. So this isn't new. This is actually, we've been doing this for a long time with scale.

>> Good job. Final question to end the segment. How excited are you around the fact that generative AI's here, and how would you explain to someone watching what's different about generative AI with Qlik going forward? What's going to be the big change in value and user experience? How would you describe what's different in this new wave of generative AI?

Drew Clarke

>> What's different is that individuals will be able to do their jobs faster, better using the trusted data that comes from a great foundation, but is combined unstructured and structured. You get the right account balance, but they also know, and then can have a conversation with you about that balance. That's what's different.

Drew Clarke

>> Yeah.

>> Drew Clarke, General Manager, Executive Vice President here on theCUBE with the data business unit at Qlik. We're bringing you all the data here live on theCUBE. Stay with us for more coverage from Orlando from Qlik Connect 24. We'll be back after this short break.