Auto-translation used
Read the original

Datanomix at the annual Qlik Connect 2025 conference in Orlando

At Datanomix, we are at the epicenter of the events of the annual Qlik Connect 2025 conference in Orlando.

AI is taking shape before our eyes. While many companies are only making plans to introduce artificial intelligence, Qlik is already showing how it works in practice, bringing tangible results.

Welcome to Agentic BI. 

One of the announcements was the launch of the new "agentic experience" Qlik, a natural language-driven interface that promises to simplify interaction with data. Instead of creating dashboards or running reports, users will be able to communicate with the system as an interlocutor to find insights, ask questions, and take action. 

In the case of Qlik, this means an entire ecosystem of specialized agents embedded in (of course) the Qlik Cloud. 

Among them:

  • Qlik Answers: First introduced for unstructured data, this tool now works with structured data as well, providing clear and reliable answers to queries formulated in ordinary language.
  • Discovery Agent: A proactive tool that continuously monitors datasets to identify anomalies or opportunities before they become apparent.
  • Pipeline Agent (early concept): It is designed to help users describe the desired business outcome, and then automatically generate the necessary data pipelines to achieve it.

Partners also enjoy the Qlik AI-agent environment.

SalesAI is a personal sales assistant that provides instant access to the knowledge and insights needed to successfully work with customers.

We ask:

  • How does using Qlik Talend Cloud for data integration eliminate data fragmentation, improving analytics and decision-making?
  • What are the advantages of Qlik over Power BI? 

And in return, we receive instant insights based on selected data from Qlik's extensive sales and marketing content, which will help us navigate the needs of customers in various industries, regions, and business challenges.

The second day of Qlik Connect 2025 was dedicated to the practical application of the latest developments and close cooperation among Qlik partners. 

Our team spent this day on the Show Floor, where we were able to see firsthand the strength of the Qlik partner ecosystem and present our own solutions.

In partnership with Motio, we demonstrated DevOps solutions for Qlik:

  • Soterre & DevJeannie - Zero-touch version control and analytics over analytics. Displays DevOps data and usage statistics of your applications directly in the Qlik interface.
  • Guided Tour - Increasing end user engagement. Provides an interactive tour of the app to understand the content and usage scenarios of the app.
  • Gitoqlok - Monitoring and visual comparison of application versions in Git. A set of tools for improving the effectiveness of AI development.
  • QSDA Pro - Automated application testing. Identifies 30+ types of errors, performance issues, and deviations from best development practices.

Continuing the AI theme: The real AI is the first BI process in action. 

At one of the Qlik trainings, they showed how an AI agent can create a data mart on its own, while being completely transparent and controlled by the user.

How it works: the developer simply sets a task for the assistant in natural language: "I want a showcase for analyzing the effectiveness of delivery, with sections by product and supplier"

What happens next: The Qlik AI agent builds the star schema on its own, selects the necessary tables, creates basic metrics, and visualizes the entire structure - data sources, fields, and relationships between them. 

The matter does not end there. The developer asks to add new dimensions — stores and warehouses. The agent finds suitable fields in available sources, explains how they relate to the already selected data, and complements the showcase scheme.

All this is done interactively, without a single line of code. You are in full control of the process: you see every step of the agent, you can intervene, reject or supplement his proposals. This is not a replacement for BI specialists, but a powerful tool that allows them to focus on really important tasks, greatly speeding up routine operations. 

The highlight of the third and last day of the conference was the public announcement of a new product in collaboration with Motio, Inc. - DevJeannie for Qlik Sense & Qlik Cloud. 

The discussions focus on the current issue of migrating analytical reporting from local Qlik servers (On-premise) to the Qlik Cloud environment. Of the Qlik customers we interviewed, about 80% are currently at various stages of this transition. In this regard, there are urgent questions: How do I prioritize report migration and how do I track actively developed applications that require special control?

The solution to these problems is offered by Qlik App Analytics from DevJeannie. This tool is an "analytics on analytics", allowing you to track in real time the number of active tasks related to a specific application, evaluate its relevance in various environments or data streams, and much more.

In between the networking sessions, an exclusive interview with Qlik CEO Mike Capone took place on the introduction of Enterprise-ready AI.

How can companies use AI more effectively in terms of costs?

Too many companies start by asking, "What is our AI strategy?" when it would be more accurate to ask, "What is our business strategy, and how can AI enhance it?". When the focus remains on results, then value comes in.

The first step is always to determine the business outcome that you are trying to achieve. If you are just starting out, it is wise to focus the initial use of AI on internal applications that will increase productivity and reduce costs, as they often provide the fastest return on investment. But then there are many cloud services and pay-as-you-go technologies that can help you control the costs of AI development, deployment, and scaling, eliminating the need for expensive on-premises infrastructure. For example, Amazon Bedrock provides access to several basic models through a single API. The main thing is to “put aside the hype” and be guided by value.

Data fragmentation has been a big issue for companies when it comes to using data for AI. How does it change?

Computing power has become much cheaper and much more scalable, and this makes a huge difference. But now we have open table formats like Apache Iceberg that allow us to structure and combine data in real time. You can quickly view and move data — which is really important — and manage that data on-site, without having to move it to a single central storage. We developed Qlik Open Lakehouse on AWS to take full advantage of Iceberg's open, real-time architecture and help companies use their data faster and more efficiently. At the same time, companies have long faced the challenges of combining information from structured and unstructured data — our intelligent interface helps solve this problem by providing smoother exploration and decision-making across both types of data.

Why do investments in quality and data management provide a higher return on investment than investments in new, larger AI models?

Data is the real differentiating factor in the age of AI, not the models themselves. The customers who have worked the most to build a solid data base are the fastest to implement generative AI. Interestingly, many of our clients from regulated industries such as financial services and healthcare are leading the way in this area because they have already invested heavily in governance, access rights, and data quality. Investing in quality and data management doesn't just improve your current AI applications; you're building a foundation that will continue to benefit no matter what models come up in the future. And with protective mechanisms in place, you actually give your organization the freedom to innovate faster and more confidently.

Everyone assumes that if you load data into a model, it will work magic. But time and time again we see catastrophic failures that occur when data is not properly monitored. This is what distinguishes AI success from failure.

402

Vitaliy Trenkenshu

May 27, 2025 11:42

Comments 0

Login to leave a comment