Qlik Connect 2025: The end of software as we know it?

A company that is hardly a household name may be about to change the software landscape for the tech giants who dominate AI conversations.

Qlik, a global leader in business intelligence and data analysis, has unveiled a bold new strategy that may rewrite the rules of the AI game.

At the Qlik Connect 2025 conference in Orlando, Florida, this week, the company introduced what it calls the “agentic experience”: a unified conversational interface that lets users tap into multiple systems through natural dialogue. The further implication is that intelligent software agents will handle everything from building pipelines to surfacing risks and recommending actions.

Rather than the kind of feature add-on that most companies bring to AI at such events, this is a direct challenge to the way enterprise software has been designed and sold for the past 30 years. It may even, if quietly and slowly, dismantle the very categories that define software.

“Today, most business software runs online, and users increasingly expect to engage with it in plain language,” said Nick Magnuson, Qlik’s head of AI, in an exclusive interview with Gadget. “But we’re moving to a point where you might not use the software directly at all. Instead, you’ll interact with intelligent agents that act on your behalf, while the software quietly powers the process in the background”.

Qlik’s architecture, called the “agentic framework”, supports a family of agents that operate within this new experience. These include Qlik Answers, which brings together structured and unstructured data in a natural language interface; a discovery agent that identifies risks and opportunities; and a pipeline agent that builds data flows based on the user’s stated business outcomes.

_{Qlik CEO Mike Capone unveils the Agentic Experience strategy at the Qlik Connect 2025 conference in Orlando, Florida, this week. Photo: ARTHUR GOLDSTUCK.}

“The most compelling part is the ability of agents to reason across steps, like a team of humans solving a problem together,” said Magnuson. “One agent can inform another, iterate, and course-correct on the fly. That’s where you move from chatbots to collaborative AI.

“The act of going to a dashboard to get your insight – that option will still be there, but over time it will become less and less the default medium by which you engage with data. It might be you’re getting insights fed to you. It might be that you are trying to find insights by engaging in natural language”.

The implications go far beyond user convenience.

Asked whether he thought this process could collapse traditional software categories, he said: “I fundamentally agree. The way in which you interact with data is going to change fundamentally. Going forward, particularly the younger generation – they’re accustomed to having things sent to them, brought to them, and not having to go and find it. And that’s the direction agents allow us to evolve the data and analytics experience”.

He said every online software platform now faced the baseline requirement of natural language interaction.

“Everyone’s got a chat icon in their application now, where they use natural language to work with that app. I do think that fundamentally is going to change how particularly online software gets built and brought to market.”

Magnuson said the software industry was not fully ready for this transformation, but vendors had no choice but to move. “The software providers are adapting. Whether their user bases are ready, I think is more of an evolution. But you have to make that pivot as a vendor, otherwise you’re considered to be antiquated.”

However, Magnuson said the agentic experience was not about replacing other software tools.

“You’re enabling more people to utilise data and analytics to inform their decision-making, because you’ve given them a way in which they can interface with the data in a more natural way and given them more flexibility in how they want to express their needs. The background agents do all the work to make it happen.

“In these architectures, agents can communicate across one another to reason through a challenge and iterate through it. The full potential of AI, not even agentic AI, is not possible without interoperability across systems, so that you don’t have to log in here.”

Charles Link, director of data and analytics at Qlik customer Reworld, a waste-to-energy business, offered a human-centric analogy: “Think of agentic AI as an intern or an apprentice you’re training. At first, you’re cautious. You don’t give them full control. But over time, they start to handle things you might not even see. Eventually, they’re good enough to take over.”

* Arthur Goldstuck is CEO of World Wide Worx and editor-in-chief of Gadget.co.za. Follow him on Bluesky on @art2gee.bsky.social.