The Next Question Matters More Than the First Answer

With Claude Dashboards entering beta, Anthropic is making it possible for people to ask quick, exploratory questions about their business data and get a dashboard back. You don't need to write SQL or build a visualization.

I think that's a good thing. We've spent years trying to make analytics more accessible, and anything that shortens the path from question to answer is progress.

I'm increasingly asked what the future of business intelligence looks like when AI can do more of the analysis for us. In fact, that was a topic with our Luminary and Partner Ambassador groups last month. My answer usually starts with a more basic question: what do people actually need from analytics?

The dashboard was never the end goal

Despite proclamations that dashboards are dead (with myself guilty as charged), they of course still persist. But people increasingly expect to get answers without going into a separate application, finding the right dashboard, and working out what it's telling them.

Picture a sales leader on Monday morning. They don't want to open five reports and figure out which one to trust before their forecast call. They want to know what changed and where to focus. If Claude makes that first step easier, I'm all for it.

Where the first answer runs out

Say revenue is down 8% in EMEA. Claude generates a dashboard showing the decline by month, product, and customer. Few sales leaders would stop there. They'd want to know whether it's one product or all of them, and whether they're losing customers or existing customers are simply spending less.

Then there's the question people tend to skip: when two people ask about revenue, are they even talking about the same thing? Sales might be looking at bookings while finance reports recognized revenue. Both numbers can be correct and still answer different questions.

Anyone who has spent time working with enterprise data knows how quickly these things get complicated. Data sits across CRM, ERP, and finance systems, and a query can be technically correct while still pointing you toward the wrong decision.

It’s great that Anthropic makes the queries behind its dashboards visible. But seeing how a number was calculated doesn't tell you whether it's the right number to act on.

And there's another consideration. What happens when the same question is asked next week, by a different person, or in a different AI tool? Will everyone get an answer based on the same definitions?

AI won't resolve disagreements about business definitions simply by generating a dashboard. If sales and finance start with different versions of revenue, they'll still get different answers. AI just makes it possible to reach those answers much faster. And at much greater volume.

For businesses, consistency matters just as much as speed. An analysis needs to be something people can explain, reproduce, and use with confidence.

The value is in the investigation

An initial dashboard can tell you revenue has declined. Understanding why, and what to do about it, takes deeper investigation.

Back to the sales example. With an agreed, governed revenue measure in Qlik, the analysis can start from the same definition finance uses. You might find that one customer segment accounts for most of the decline while another is growing. Or that the problem isn't regional at all, but a product category performing differently across markets.

That's where combining AI reasoning with Qlik's analytical capabilities gets interesting. Claude makes it easier to ask questions and work through what to investigate next. Qlik brings governed data, business definitions, and associative analytics that help people explore the relationships behind the numbers.

And this is increasingly important as AI opens analytics to people who aren't data specialists. Making it easier to ask questions is valuable, but people also need to understand what they're being told and have confidence in the decisions they make.

It's all about choice

We know customers don't work in a single application. That's why we've invested in MCP, opening up Qlik's data and analytical capabilities to external AI assistants. The new native Qlik connector within Claude takes this a step further, making it easier for customers to access Qlik directly from an AI experience they're already using.

This isn't about choosing between Qlik's own analytics and AI experiences, including Qlik Answers, and assistants like Claude.

A customer might start by asking Claude a question, then move into Qlik to explore the relationships in more depth, before returning to the AI assistant to investigate something else.

We need to make that experience as connected as possible. People shouldn't have to think about which application they're using every time they want to ask another question. They should be able to work with the data and analytical capabilities they need, in the way that makes the most sense for them.

For Qlik, that means extending the value of the investments customers have already made in our data and analytics, while opening up new ways to use those capabilities.

What is the future of BI in an AI world?

I don't think the future of BI comes down to whether dashboards survive or AI assistants replace analytics applications. Most organizations will use both, depending on what they're trying to achieve.

But developments like this do change what customers should expect from analytics vendors, including Qlik.

It's no longer enough to make information available inside an application. We need to help customers work with consistent definitions, explore their data more deeply, and make those analytical capabilities available wherever they're working.

That's where Qlik's combination of data integration, governance, and associative analytics matters. We're not just helping customers access an answer. We're helping them investigate it, understand where it came from, and decide what to do next.

And through MCP and our work with Anthropic, we're extending those capabilities into new AI experiences.

So yes, I welcome Claude Dashboards. Anthropic itself positions it for quick, exploratory questions alongside established BI and analytics tools, and I think that's the right approach.

Making the first answer easier to get is progress. But I'm much more interested in what happens when someone asks the next question.

Further resources:

• Qlik MCP: Quick Demo – See how Claude can interact with Qlik's analytical capabilities through MCP, using natural-language questions.

• Connecting Claude to Qlik MCP – A practical walkthrough of connecting Claude to Qlik.

• Qlik MCP Server: Overview and capabilities – Explore how Qlik makes its governed data and analytics capabilities accessible to AI assistants.