Share Your Experience

[SarahUrbiss]

What's your experience? Share with the Community and earn the **new** Agentic AI Expert badge.

You've moved past the first try. Now tell us what happened when agentic AI stopped being an experiment and started being part of how you actually work.

Post your experience here and we'll award you the Agentic AI Expert badge.

Tell us in the comments:

• What result or takeaway can you share?
• Do you have a metric or success story? For example, time saved, faster answers, fewer manual steps?
• How did you (or how do you plan to) expand this beyond a one-off experiment into a regular workflow?
• Got a screenshot? Share it if you can. Seeing the real thing helps everyone else picture it in their own environment.

Bonus: tag a colleague who should be trying this too. By sharing your experience, you help inspire others and generate ideas in the Community.

Everyone who shares their experience here gets the Agentic AI Expert badge.

Thank you for taking the time to share!

[Chanty4u]

Hi @SarahUrbiss it's nice to see these kind of ideas..below is mine very simple 

My experience with Agentic AI

After exploring Qlik’s Agentic AI capabilities, my biggest takeaway is that the real value is not just getting an answer in natural language—it’s reducing the number of manual steps between a business question and an actionable insight.

One workflow I’m exploring is using Qlik Answers for natural-language analysis of business data. Instead of creating a new sheet or writing expressions for every question, users can ask questions such as:

- “What are the top-performing customers?”

- “Which region has the biggest decline?”

- “What changed compared with the previous period?”

- “Are there any unusual trends I should investigate?”

This can make self-service analytics much easier, especially for business users who don't want to navigate through multiple dashboards.

My key takeaway: Start with a focused business use case, validate the AI-generated insights against trusted Qlik data, and then gradually expand the workflow.

The next step for me is to move from a simple Q&A experience toward a more repeatable agentic workflow, where AI helps identify insights and guide users toward the right analysis or action.

Best

Chanty4u 

[QFabian]

Hi @SarahUrbiss !, Hi @Chanty4u ! 

Using our existing Qlik Apps as the core data and logic foundation for Qlik Answers has been the single most effective lever for driving smooth, reliable Agentic AI adoption across our organization.

Instead of building AI agents in a vacuum, grounded context is everything. By bridging our validated Qlik data models with Qlik Answers, the AI agents leverage business logic, governance, and existing metrics.

Users already know how to make a good prompt, so knowing Master Items, users can ask questions using our standardized business terminology and trust that the agent evaluates metrics using the exact calculation logic defined in our Master Items—completely eliminating field-name guesswork and hallucinated KPI logic.

[MatheusC]

Hi, @SarahUrbiss 

When we started using Qlik Answers in our organization, we quickly realized the importance of preparing our data structure in advance. We adjusted field names so that each one could be easily understood, standardized terminologies, and created master items. This initial preparation allowed users to experience *simplicity, speed, and confidence* in the answers provided by Qlik Answers.

What I would like to share as my key takeaway is this: before anything else, review your data. Are they clear, clean, and consistent? Today, Qlik Answers also offers *AI-powered optimization*, which automatically provides summaries and detailed views of your fields and master items. You can further refine this view by applying the necessary configurations. It is indeed a challenge, but once solved, you will see how much the experience improves and how powerful the tool becomes in everyday use.

Regards, Matheus!

[robert_mika]

I connected my dataset of the highest UK mountains to Qlik Answers in Qlik Cloud, and what started as a simple experiment quickly turned into my default way of exploring data.

Instead of manually building charts or setting up complex filters, I now just ask Qlik Answers direct questions in plain English. For example, when I asked it to summarize the peak values and elevation trends, it instantly pulled together the key metrics, averages, and height distributions without me having to slice the data myself.

The biggest takeaway is how much time it saves by eliminating manual steps. I save about fifteen minutes of dashboard configuration every single time I need quick insights on peak heights or regional stats, getting instant answers instead.

To make this a regular workflow rather than a one-off test, I set up the business logic in Qlik Cloud so the AI naturally understands geospatial terms, regional names, and mountain metrics. Going forward, I plan to upload route guides and PDFs into a knowledge base so I can query trail info and mountain data all in one conversational spot.

[ssomaiNY1]

I've broken down the six agents behind the Qlik development workspace

- App Inspector.
- Script Writer.
- Script Reviewer.
- Front-end Builder.
- Architect.
- Tenant Administrator.

Individually, none of them is especially revolutionary. The interesting part for me is what happens between them.

- An inspector can understand an app without changing it.
- A writer can create a script without deploying it.
- A reviewer can challenge the writer without rewriting the work.
- A builder can change the application, but only within an explicit approval.
- An architect can challenge the design.
- An administrator can investigate the platform much more broadly than it is allowed to change it.

And underneath all of them sits the same rulebook.That is probably the biggest lesson I've taken from this experiment.I started out thinking I was building agents.I ended up spending much more time designing responsibilities, controls and hand-offs.

I'll share how I assembled it rather than pretending there's one perfect configuration everyone should copy.  Qlik Agents  


If you were building your own version, which role would you add, remove or combine?

[juanfrantamaral]

We connected Claude to Qlik Cloud through the Qlik MCP Server, so our sales team can ask questions in natural language from an AI tool they already use, with no need to open an app and apply selections.

The MCP server reuses the governed apps and data models we already have. Same business definitions, same Section Access, no new pipelines. Clean, simple data models (no synthetic keys, no snowflakes) made answers reliable from day one.

We started with a narrow use case, quick customer and product lookups for sales. Next we'll add more apps and other teams, and complement it with Qlik Answers and Discovery Agent for proactive insights.

[hesh]

I recently explored Qlik Discovery Agent using a retail RainWear monitoring scenario inspired by a client engagement (business labels anonymized).

The agent proactively identified:

• Record high sales during monsoon season.
• Record low sales during winter months.
• Higher-than-expected sales performance.
• Regional sales downtrends requiring attention.

What stood out was how the agent continuously observed data, detected anomalies, generated business-friendly narratives, and guided further investigation.

Attached are a few screenshots from the exercise. Happy to hear how others are using Discovery Agent in their environments.

More details in my full write-up here: 

https://community.qlik.com/t5/Put-Agentic-AI-to-Work/Putting-Agentic-AI-to-Work-with-Qlik-Discovery-...