Trusted data key for Qlik as it develops foundation for AI

Qlik aims to be the trusted foundation for AI.

Data is what informs AI models and applications, including generative AI. It's what gives AI models and applications their knowledge so they can deliver useful outputs that inform decisions. Meanwhile, sparked by OpenAI's launch of ChatGPT in November 2022, interest has surged in both traditional AI as well as generative AI.

Driven by that swell of interest in generative AI over the past 18 months, AI has become the dominant trend in data management and analytics, as well as most other areas of tech.

Vendors want to develop AI-powered capabilities such as natural language assistants, text-to-code translation capabilities and environments for customers to create their own AI models and applications. Enterprises, meanwhile, want to take advantage of such capabilities, training AI models with proprietary data so the AI assets understand their business and can simplify data management and analysis.

Tech giants AWS, Google and Microsoft are among those that have prioritized developing tools powered by generative AI, such as AI assistants, as well as environments that enable customers to create their own generative AI applications. So are data management and analytics specialists such as Informatica, MicroStrategy and Tableau.

Qlik is right there with them.

Based in King of Prussia, Penn., the vendor is a longtime analytics specialist. In recent years, Qlik added a competitive data integration platform and developed AI capabilities such as natural language processing and automated machine learning.

Using those AI, BI and data integration capabilities as a base for creating an AI development environment, one of Qlik's main goals is to provide customers with a trusted foundation for building new models and applications, according to Mike Capone, the vendor's CEO.

Mike Capone

Toward that end, Qlik in September 2023 unveiled Staige, a portfolio that brings together the vendor's AI and machine learning features in a single location and combines them with generative AI capabilities from third-party large language model vendors such as OpenAI and LLM platform provider Hugging Face, among others. In addition, while Qlik historically only provided support for structured data, in January 2024 the vendor acquired Kyndi to add support for unstructured data.

Together, the vendor's data integration capabilities that help ensure data quality, analytics capabilities that enable accurate analysis and Staige -- which will include Kyndi's capabilities -- provide customers with trusted data they need for AI models and applications, according to Capone.

With Qlik Connect, the vendor's user conference in Orlando, Fla., scheduled for June 3-5, Capone recently discussed Qlik's AI strategy.

He spoke about what the vendor has already done to foster generative AI development as well as what new generative AI-related technology it plans to reveal during the conference. In addition, he touched on the state of Qlik's construction of an environment for AI development and the challenges customers face as they attempt to build trusted AI models and applications.

Before we get into specifics about Qlik, in general, how do you see generative AI changing business -- what will it enable enterprises to do with data that they couldn't in the past?

Mike Capone: GenAI is going to be transformative. It's going to change the way we think about and interact with data. It's going to transform businesses, embassies, governments -- it's going to change society. That being said, you can't skip to the end. You can't just plug in GenAI and have everything work. What we're focusing on is data mastery, which is getting the data that's underpinning GenAI to be of the quality that you can actually trust what's coming out of GenAI.

You can't skip to the end. You can't just plug in GenAI and have everything work. What we're focusing on is data mastery, which is getting the data that's underpinning GenAI to be of the quality that you can actually trust what's coming out of GenAI.

There's the old expression, 'garbage in, garbage out,' and we now have the ability to spit out garbage at tremendous velocity. With the new computing power we have combined with LLMs, if the data quality is bad, you'll not only get bad answers but you'll get them really fast. What we're focused on is harnessing data in an honest, ethical way while ensuring that not only is the data available but also the quality of it is to the highest level so that AI can do its thing.

As you've shaped Qlik's generative AI plans over the past 18 months or so, what is your vision for what Qlik will ultimately provide customers?

Capone: We want to be the trusted foundation, the trusted platform, on top of which customers can build AI capabilities. We believe that large language models and other mechanisms will evolve and people will switch and change between them. But what you can't do without is data quality, and you can't keep redoing the foundation. You have to get that right. We will have tools and capabilities that not only work inside our platform but also interoperate with other tools in a very open fashion so we can provide that [data quality] to ensure an organization's AI deployment is effective.

As we talk some 18 months after generative AI became the dominant trend in analytics and data management, where is Qlik on its journey toward providing both generative AI capabilities as well as an environment for customers to build trusted generative AI models and applications?

Capone: Where we are is that we've arrived. We already had a big part of [the foundation]. We acquired a company called Attunity in 2019 that enabled us to harness data at high velocity in real time from any source and then put it in any target. Getting the data is a big part of it. But the piece that we didn't have was the data transformation, data quality, data trust capabilities. Last May, we closed on our acquisition of Talend and now we're a year into it and have fully integrated those capabilities into our platform. Now we have data integration, trust, quality and transformation capabilities plus the 30 years of experience in analytics, natural language processing, automation and AI that we bring to the table.

My message when we get to Qlik Connect in Orlando is that we're there.

As we speak a few days in advance of Qlik Connect, what specific generative AI-related capabilities has Qlik provided so far?

Capone: We've had a pretty steady drumbeat of AI announcements. Our platform name for our AI capabilities is Qlik Staige. It encompasses all our GenAI capabilities. On the data integration side there are embedded AI capabilities to look at data quality as data is being transformed. On the analytics side, we have automated machine learning capabilities to take the data science world that has been isolated and bring that mainline into analytics. We've always had something called Insight Advisor, which is our AI and natural language processing agent on top of structured data. Now, after our acquisition of Kyndi, we're adding unstructured data, which is super important because 80% of data is unstructured. But we haven't formally launched that product.

What can customers expect to hear next week related to AI and generative AI?

Capone: What they're going to hear is that we're going to cut through the hype. A year ago, it was a free-for-all. [The attitude] was 'just get AI going.' Sanity has returned. There's no skipping to the end. What we're going to talk about is that we are a trusted partner in the foundational space to underpin AI. We're going to talk about the importance of governance, data sovereignty and how our platform enables users to manage those things in a secure fashion.

We're also going to talk about our commitment to being open. Our firm belief is that you shouldn't be in a position where you have to give all of your data to somebody to make AI work. You should have choices. You should also be able to harness your data where it is and not have to move it to somebody else. We don't have to own that piece of it. We have to own the outcome and make people successful.

Regarding openness, does Qlik integrate with a variety of LLMs and database providers?

Capone: We tend to keep up with customers' demands and we will have APIs out to everybody. For example, we have an API out to ChatGPT so that if a customer has any QlikSense application in our cloud -- the work has already been done to curate the data -- they can then snap ChatGPT on top of it and start interacting with it in the ChatGPT modality.

As this new era evolves, what are the biggest challenges vendors and customers face as they try to operationalize generative AI?

Capone: There's a stat that 72% of organizations are being constrained in their AI work by not being able to correctly manage their data. Underneath that, trust and quality come up in all my conversations. It's really hard to build trust, but it's really easy to destroy it. We saw this with vendors years back claiming they could read X-rays better than doctors or make better diagnoses, and it set the industry back. The same thing can happen inside a company if someone rolls out a GenAI companion and the first three answers they get are ridiculous. They won't want to use it anymore. That trust factor is really important, and that's why the curation and quality of the data are critical.

People are saying, 'Help me. Help me make sure I don't mess this up by putting out things that turn people off within my organization, or worse, land me in trouble with regulators because I did something I shouldn't have done.'

How is Qlik addressing those challenges?

Capone: Our platform is basically purpose-built to govern your data to make it ready for AI, everything that helps with data quality. We're building trust scores around AI data quality. We have data lineage capabilities built into our platform so you can see whether data belongs in an application. Those are some of the things we're doing to give customers assurance. And it's all governed by our platform so all the capabilities around security are there.

Editor's note: This Q&A has been edited for clarity and conciseness.

Eric Avidon is a senior news writer for TechTarget Editorial and a journalist with more than 25 years of experience. He covers analytics and data management.