Qlik Connect 2026 | Nick Magnuson, Qlik & Mike Leone, Moor Insights & Strategy

Clips
More from Qlik Connect 2026

Nick Magnuson

Head of AI

Qlik

Mike Leone

VP & Principal Analyst

Moor Insights & Strategy

play_circle_outline AI enabling autonomous actions, not just insights, changing data lifecycle requirements

play_circle_outline AI-driven shift transforming analytics, data engagement, and decision-making paradigms

play_circle_outline AI Slashes Time-to-Insight by Linking Data — Data Quality Is Key to Avoid Hallucinations

play_circle_outline Closing the Pilot-to-Production Gap: Governance, Observability, and Guardrails for Trustworthy Enterprise Deployment

Info
Transcript

Nick Magnuson, Qlik & Mike Leone, Moor Insights & Strategy

Nick Magnuson

Head of AI Qlik

Mike Leone

VP & Principal Analyst Moor Insights & Strategy

In this interview from Qlik Connect 2026, Nick Magnuson, head of AI at Qlik, joins Michael Leone, vice president and principal analyst at Moor Insights & Strategy, to talk with theCUBE Research's Rob Strechay and Rebecca Knight about how enterprises are moving from dashboard-driven insights to AI-powered automated decisions. Magnuson outlines how the traditional model of building reports and acting on them manually is being overtaken by AI systems capable of detecting patterns, analyzing data at scale and taking action autonomously. Leone highlights that data... Read more

explore Keep Exploring

How is AI changing the way organizations engage with, manage, and act on their data, and what implications does that have for existing data frameworks and lifecycle practices? add

How should data platforms, architectures, and frameworks evolve to support autonomous (agentic) systems and accelerate time-to-insight—while addressing data quality and using patterns like Iceberg and the Medallion approach to avoid data duplication? add

How is AI affecting analytics—particularly time-to-insight, data discovery and quality—and how should we weigh concerns about AI hallucinations versus human error in decision-making? add

What are the main reasons AI pilots fail to scale from pilot to organization-wide production? add

bolt Powered by CUBE AI

Nick Magnuson, Qlik & Mike Leone, Moor Insights & Strategy

search

Rebecca Knight

>> Hello, everyone. And welcome back to theCUBE's live coverage of Qlik Connect 2026 here in Orlando, Florida. I'm your host, Rebecca Knight, alongside Rob Strechay, my esteemed colleague. We're getting on with this show.

Rob Strechay

>> It's great. I think the amount of information that we're bringing to demystify a lot of what goes on between the data and the AI and agentic has really been great.

Rebecca Knight

>> Well, and here to talk more about that, I would like to welcome two guests to the show, Nick Magnuson, head of AI at Qlik. Welcome, Nick.

Nick Magnuson

>> Oh, thank you for being here. Yeah.

Rebecca Knight

>> And Mike Leone, VP and principal analyst at Moor Insights & Strategies. Welcome.

Mike Leone

>> Yeah, thanks for having me.

Rebecca Knight

>> Yeah.

Rob Strechay

>> First off, congratulations on the new title and everything, as well.

Mike Leone

>> Oh, thank you.

Rebecca Knight

>> Yes, very good. I know, these analysts, they just have to-

Rob Strechay

>> Yeah, well. We've known each other a long time.

Mike Leone

>> Yeah.

Rebecca Knight

>> So I want to start with you, Nick, and maybe just talk at a high level, a broad level, about this shift that's happening right now in analytics and what is driving it.

Nick Magnuson

>> Well, obviously AI is playing a big role and that shift is really changing the way we think about how we engage with our data. Traditionally it was, "Hey, we're going to build a dashboard, we got reports, we're going to act on those things," but now AI can do a lot of that stuff on our behalf and even act on our behalf. So it's changing a lot of paradigms from how we think about acting on that information, how we think about constructing data and supporting that data through that life cycle now, because we've got autonomous things in the mix that aren't human in nature. So I think a lot of the frameworks that we've used in the past now need to be rethought essentially from the ground up.

Rob Strechay

>> Yeah. So, Mike, we've been around the block for a long time on this stuff. And when you start to look at this, a lot of it was, "Okay, how do we connect the data up?" It was a lot of good stuff around Iceberg and a number of other ways of not moving the data around so we're not duplicating it and we're kind of going through the Medallion strategy. How do you see this really having a platform that connects all of this together, really aiding and getting to agentic?

Mike Leone

>> Yeah. And just speaking on what Nick had just mentioned, I think AI has enabled us to really solve that time to insight metric. That's been what's driving analytics for so long. And I think we've all heard the garbage in, garbage out. We've all heard the issues around data quality. And that was a problem well before AI entered. I think the enablement that we're going to get from AI now is, "Hey, I don't have to rely on a human to try and go figure out where to pick my data that I'm going to be using to analyze." It's going to find it pretty darn quick and it's going to do it faster and it's probably going to do it more accurately than what I think humans have been able to do. One of the things that I think is overlooked, we hear so much about AI hallucinations when it comes to arriving at a decision. I'll guarantee you, if you go and ask any CEO, any management structure, "Have you had a human in your team that has made a bad decision, has arrived at the wrong insight?" They're going to say yes. So I think there's some components there of like AI helping to leverage more data to act on insight. I think it's going to be beneficial for a lot of organizations.

Rebecca Knight

>> So how are you seeing companies pressure test that though in terms of really where the hallucinations come from, the bad data? How are you seeing companies pressure test that? The data that their systems are relying on.

Mike Leone

>> Yeah, you know what? It's funny, right? I think we've heard so much about this pilot purgatory and going from pilot to production and you hit production and then all of a sudden everything goes haywire. And when you think about those pilots, it's, "I have this perfectly curated data set and I have these folks that are testing this out that have a high technical IQ. They know AI, they know its limitations, they know how to prompt." And then all of a sudden you say, "Hey, here we go. Let's give this to the entire organization." And you see it fall flat on its face. And I think a lot of the things that we heard today in the keynote, the focus on trust and transparency, all the work that Nick and his team have been doing to improve accuracy, I think that that goes a long way. And I think that's truly the barrier to success. It's transparency, it's trust, it's security, it's privacy, it's governance, guardrails, all of those things. So the heavy focus that we're seeing from a lot of vendors, Qlik included, on that observability piece, I think is going to go a long way in solving some of the problems and ensuring a higher level of trust.

Rob Strechay

>> Yeah. I think, again, when you start to look at how all of the pieces are coming together, and really in particular I look at it as plugging the models in is the easy part. And when you start to look at it, the models -- I loved the con -- you're going to switch out models, you're going to switch out agents. Nick, from the perspective as you're building AI -- big word AI, right? -- so when you start to look at that, what are the key challenges that you're seeing about connecting the full data estate in that? Because I loved the comments about data products and semantics down through open standards and open semantic layers and things like that. But connecting that all together is still, to make it make sense is not easy.

Nick Magnuson

>> Yeah. It's not easy, but that's where the value comes from is if you can bring that full data state. Like AI is better at humans at finding patterns in data that seems interrelated but not connected and maybe there's no pattern in it. AI can find that better than us. And then it can do it at scale and at speed. So you're right, making sure that you connect that data, you create the interfaces between the different connection points, and then provide that metadata layer, that semantic layer, so that when these things are operational, there's the connective tissue that allows you to get the true value out of it. At Qlik, we spend a lot of time thinking about how do we connect and integrate across the entire AI landscape. Wherever your data is, as we expressed today, that's your decision and we're going to support it and we'll integrate wherever. That way the data doesn't move, you don't have to go into a single stack situation, and you can remain independent, and as we talked about today, you have freedom. You have freedom to change, you have freedom to move as the technology is going to change. That's the one thing we know is the technology moves faster than it's ever moved before and you have to have the ability to be agile and move in the new directions and continue to leverage that data state that you've invested in for many decades.

Rebecca Knight

>> So you say that AI is better at humans at detecting patterns and recognizing, seeing signals better than us, it could do it at scale. As the head of AI, what do humans do better than AI?

Nick Magnuson

>> Well, humans are better at devising the system in which those things can operate, right?

Rebecca Knight

>> Okay.

Nick Magnuson

>> And that's the good thing. We have that superpower that we kind of sit above it, and the ability to put in the governance frameworks and the things that make it work over time. AI is not a point in time, it's an over time situation. So you've got to be able to have humans that can get in there and assess the thing, put together the system, and then monitor it. And it may be a situation where you got to inject new data, you got to inject new context we talked a lot about today. That's a big thing. If models don't perform well, you got to have governance. And when the governance tells you something's off, you got to provide some human intuition back into it to reinforce it so that you can use it to operate as you expected.

Rob Strechay

>> Yeah. I mean, I thought that the whole comment around the chief context officer was very ... I mean, to me, it would seem that it's going to become everybody's role, right? I mean, when you look at AI, it's about the process and the use case and the context of which the data. Do you see this as forming how you're building the platform? And how are you seeing it, Mike, from a perspective of organizations out there really absorbing and who owns the context about that? So I'll let you hit it first from-

Nick Magnuson

>> Yeah. I think context is a new term that we're all trying to understand what it really means. I can tell you context is king when it comes to LLMs and using AI. Models are ... They're like you're training a puppy. You put the food out there and they'll do whatever trick you taught them last, even though you're trying to teach them a new one. They're eager to fill in the gaps essentially. So when trusted data's not there, when you got gaps in your data, they're going to hallucinate because they want to appease the human there. So context is super important. I think one of the things we didn't talk about enough today on stage is context is not temporal, meaning it should evolve and grow with you as an organization and as a user over time so that as you continue to operate with AI systems, they can lean on things that you've done in the past and apply them to things in the future.

Rebecca Knight

>> So that they have their own institutional knowledge.

Nick Magnuson

>> Correct.

Rebecca Knight

>> But that's what context is. Yeah.

Rob Strechay

>> And how are you seeing it from an organizational ... Because you're talking to CIOs and CDOs and things like that. No chief context officers yet, so.

Mike Leone

>> Yeah, I don't know if that's going to be a title, like you said. I think everybody's going to kind of be in charge of context and how it applies to their specific use case. I think the folks that are asking LLMs or asking agents to do work on their behalf and to take action, they kind of have an idea of what they want to do. They're not coming in completely blind. I think it's been really interesting. I'd say over the last year and a half or so, we've started to see a reemergence of what a semantic layer means. And we saw that today on stage, and I love some of the things that I was hearing on stage, but that's a big deal. You have to be able to understand what's happening and connecting it to the business. And I don't want to get too technical, but we talk about knowledge graphs and connecting some of those dependencies that maybe you didn't know existed before. That goes back to what we were talking about earlier around, "Hey, what can AI do better than humans?" It's understanding relationships and bringing that back as context to make informed answers, like deliver informed responses. So I'm really excited to see the reemergence of that semantic layer and how valuable I believe that that's going to be to generative AI and agentic AI.

Rebecca Knight

>> Mike, how exposed are companies right now when it comes to the explainability and defensibility of AI driven decisions?

Mike Leone

>> Oh, boy. Very.

Rebecca Knight

>> No one wants to be a headline.

Mike Leone

>> Yeah. Well, that's it. And Nick and I actually talked about this earlier today. I think there's some companies that can handle being the headline. I think a vast majority of organizations can't. And I think that's where the hesitancy comes in. I think that's where needing more trust in not just the data, that's important, but trust in the decision or the action that's being taken based on the data. So it's tricky right now. I think that if you survey 10 folks here, I think you'll see two of them that are going to lean heavy on the innovation side and say, "Hey, you know what? I'll take on the risk of potentially running into a problem because I want to be a first mover. I want to announce this new capability." I think you're going to see a vast number of organizations that fall into that, "It really depends on what we're doing here." And then you're going to see folks of, "I don't trust this yet. I'm not even going to go down this path. I need to see more proof points." And frankly, I think the customer stories that we hear at all of these events, I think that matters a lot more than people think. I think understanding, "Hey, there's this other company, maybe in my industry, that's doing something cool and unique and they've brought this to market and it is being trusted. It is working. Maybe we can embark on our own journey now." I think that matters a lot. So I'm excited to talk to more of them while we're here.

Rob Strechay

>> So when you look at it, AI means more than just chat and LLMs and things of that nature. There's that was AI now called ML and gone back to its ... So there's a lot of different things and an agent isn't going to just be using an LLM. How do you help people from an organizational perspective with the platform really bridge those multimodal types of agentic capabilities?

Nick Magnuson

>> Well, I think to begin with, it's a journey. Every customer's going to go on their own journey of how they get to the state of where they're using agentic to really drive business outcomes. We talk a lot about like, "Hey, we went live with agentic," or, "We're now measuring adoption and usage and stuff like that." That's not really the goal when you're using agentic. It's to get to a point where you're measurably impacting operations in an improved way. And everybody's agentic journey is going to be a little different. I like to suggest that people start with something small, something that they can contain and start to build some strength and muscle and confidence in how they approach it, and then take on bigger and bigger projects. One of the things I've seen, and this goes back to my days in ML, people wanted to boil the ocean. They tried to tackle these huge problems. They uncovered all these data quality problems that they couldn't solve for, and those things got shut down. And so it's really about, let's start with something small. How can we use an agent to get something really effective, but not like try something so complicated, or if we get it wrong, to Mike's point, we're a headline risk now. That's the type of things that I suggest customers go into is something that they can measure and something that they can build from through success.

Rebecca Knight

>> How do you think about measuring it? Because as you said, the goal is to measurably impact business operations, but you're also talking about starting small, doing an experiment. So what are your best practices and recommendations here?

Nick Magnuson

>> Yeah. Well, measurement and impact, they don't have to be sized accordingly or similarly, right? So just because it's a small project doesn't mean it can't have impact. And obviously you go from there. But I think at the beginning, to answer your question, you've got to set out the expectation of what you're trying to achieve with this particular solution. And it's important, I think, from a sponsorship and success standpoint to get to the point where you can measure that ROI, is that you get the right people involved, you have the right support, you have the right ability to action and operationalize those outcomes. Otherwise, it's just going to be an experiment that dies and fails. So my perspective, yeah, focus on something that you can measure, set out those expectations, and you got to bring people into the mix to make sure that they sponsor it and, again, can operationalize that.

Rob Strechay

>> And maybe to both of you, how do you see organizations approaching it where they have to balance performance and innovation, cost? There's like kind of three legs of the stool, but you can only really have two of them in good working condition and you can't be very innovative and very performant and really cost less kind of thing. So how do you see that balance happening for organizations?

Mike Leone

>> I guess I'll go first and touch on it a little bit. One of the things I'm seeing from a lot of organizations, the budget's there, they want to make that investment and they're willing to. And I want to go back to the ROI story that you just had. I think it's okay for executive teams to not have that metric to measure complete success. I think understanding, yeah, folks are using this tool, we're seeing some productivity and efficiency gains. That's good for now. It's okay if you can't establish this amazing ROI metric, they just have to get in because they can't risk falling behind. And a lot of executive teams know that. So I know I took that question a little different there, but when it comes to performance, it's going to be a balance. And I think the vendors, the Qliks of the world, recognize that balancing cost and performance and risk, that's tricky. So the folks that you're going to want to strategically partner with here are the ones that are trying to abstract the complexity away, saying, "Hey, we want to lower the barrier to entry to using these tools. We want you to try to use these tools. We want you to see the potential value." And frankly, I think -- I mean, we talked about it a little today, right? -- I think Qlik is doing a really good job of being able to give that level of enablement to customers to say, "All right, I'm not going to break the bank here. I am going to be able to understand how these agents interact and work behind the scenes." And I think that that's going to actually broaden adoption a lot faster than people expect.

Nick Magnuson

>> Yeah. And the one thing I'd add is that it's a big ecosystem. And if you think that you're going to do all things to enable AI, you're wrong. You've got to be able to pick and choose where you have strength, focus on that, and then bring partners in. Whether you're a vendor like we are at Qlik or whether you're a customer, you've got to know where your differentiation lies, where your core IP lies, and then bring other solutions around to bring together a solution around AI. Because again, trying to do it all is unrealistic. And if you do try and do it, that's where the time element comes in and you will fall behind.

Rebecca Knight

>> But AI is not the answer to every question.

Nick Magnuson

>> Correct.

Rebecca Knight

>> Okay, all right. Profound. And Nick and Mike, thank you both so much. I really appreciate it. A very fun conversation.

Mike Leone

>> Yeah. Thanks so much for having me.

Nick Magnuson

>> Thank you.

Rebecca Knight

>> I'm Rebecca Knight for Rob Strechay. Stay tuned for more of theCUBE's live coverage of Qlik Connect 2026. You're watching theCUBE, the leader in enterprise tech news and analysis.