Death by 10000 dashboards
The future of dashboards is about how storytelling meets analysis in a way that drives action.
Dashboards were supposed to be our saviors. They promised the professionalism of a plane cockpit: clean gauges, clear dials, a single view of what’s going on. One glance and you’d know if your product or business was flying smoothly.
That was the dream.
The reality is a world apart: the feeling of being trapped in a control room with 10,000 blinking screens. Instead of clarity, we’ve ended up with chaos.
How We Got Here
Why did dashboards explode in the first place?
Data became a religion. “Data-driven” became the battle cry, and dashboards were the sermon. If you didn’t have a dashboard to measure your progress, did your work even count?
Self-service tools took over. Looker, Tableau, Power BI — suddenly anyone could build a dashboard in an afternoon. And they did. Lots of them.
Metrics fractured. Marketing’s “active user” wasn’t the same as product’s “active user.” Preferring to move fast rather than pay the price of alignment, teams created their own dashboards to reflect their own definitions.
Tools sprawled. Every SaaS tool now ships with its own analytics, so teams end up checking multiple dashboards inside and outside of their core stack.
BI became a treadmill. BI teams sprang up with the core job of building dashboards for business users. Their success naturally became measured in output (more dashboards!) rather than driving clarity or outcomes.
Dashboard bloat. Once created, they linger forever, piling up like stacks of paper on a cluttered desk.
What was meant to be empowering has become overwhelming.
The Problems
Here’s what life in dashboard hell actually looks like:
Conflicting truths. You open five dashboards and get five different numbers for the same metric. Which one do you believe?
Endless scavenger hunts. To answer a simple question, you have to stitch together views from multiple dashboards. If you want to understand whether weekly active users are growing among U.S. iOS users who are male and between 13–17 years old, you might need four separate dashboards. There’s no fast way to drill down.
Surface, not depth. Dashboards often tell you what happened, but not why. Metrics go up or down, but the root cause stays hidden.
Cognitive overload. The sheer number of dashboards creates noise. Teams spend more time trawling for useful signals than making decisions.
Lack of trust. With no single source of truth, dashboards stop being instruments of confidence and start being sources of debate.
Dashboards were supposed to help us spotlight what’s important. Too often, they simply create more fog.
The Way Forward
The answer isn’t “just one more dashboard;” it’s a reinvention of how insights and analysis are told, not just shown. Dashboards need to stop being static grids of KPIs and start becoming living narratives that answer questions.
Agentic workflows combined with rich semantic layer and deep science is crucial. A modern approach must do more than plot numbers on a screen. It should connect to a single source of truth, carry a rich semantic layer that encodes the business context, and use built-in science to deliver deeper insights. It should also feature agentic flows that don’t just present data but actively answer the questions that builders have in real time.
There are only a finite set of analysis to automate. There are, in fact, probably fewer than a hundred common types of analysis in most organizations — things like diagnosing the root cause of a metric change, running an A/B test, discovering an “aha” moment, diving into churn, or performing descriptive growth analysis. Analysts have developed manual approaches for each of these, and those approaches can be encoded. Using a combination of the determinism that algorithms and code provide with the flexibility of generative AI, we can automate these analyses so they’re done in minutes rather than days. When you combine that level of automated depth with good storytelling, the experience becomes truly magical.
The future of dashboards is about how storytelling meets analysis in a way that drives action. The power of insight lies in its ability to resonate; data only matters if it moves someone to act. Future work must give data-informed stories pulse and presence through engaging, intuitive, and personalized forms. Whether that’s a compelling slide deck for an executive, a New York Times-style narrative that walks through a trend like a story, an interactive animation for exploring “what if” scenarios, or even a dense but crystal-clear table for an analyst who wants the raw detail.
Each story should be tailored to the decision-maker it’s meant for. And crucially, the story should evolve dynamically. As the analysis deepens or the situation changes, the narrative changes with it. Updating, reframing, and highlighting what matters most in that moment!
A Glimpse of This Future
Imagine you’re a product manager logging in on Monday morning. Instead of opening four dashboards to see what’s happening with U.S. iOS male users aged 13–17, an intelligent system greets you with a living story:
“Weekly active users grew 8% week-over-week in your target demographic. The increase is disproportionately manifested among US male teenaged users due to a recent product change that was made. Here are the three levers you can pull to sustain the growth.”
If you’re an executive, the system shows you a concise narrative with key risks and opportunities. If you’re an analyst, it gives you a detailed breakdown and links to the underlying data. The content adapts to you. As the underlying data shifts, the story shifts with it.
That’s not a dashboard. That’s an adaptive, story-driven interface that surfaces the right insight at the right time, in the right form.
We can’t live in a world where every decision requires spelunking through a dozen dashboards of dubious trustworthiness. That’s death by 10,000 dashboards.
It’s time to leave dashboard clutter behind and step into an AI world where intelligence is dynamic, contextual, delightful and actionable.
My team and I were just talking about this subject last week!Building a BI dashboard in today’s age feels like a giant missed opportunity. I see a future where you can toggle modes depending on what you need in that moment - a narrative recap, infographic, or raw data, etc. There is magic in combining deterministic flows with generative AI. It reminds me of how I felt about surveys back in the day… I’d trade 50% of the responses for ONE follow up question. What if a reporting product used insights as starting points for a “data-insights conversation” and ultimately helped people ask better questions!
Excellent article, this really resonated. You've eloquently articulated a frustration that has plagued the data world for years. The move towards agentic workflows feels like the most promising path forward I've seen.
I especially loved your section on "storytelling meets analysis." This is a critical insight. Many analysts and engineers can find the 'what', but struggle to build a compelling narrative around it. Using AI to bridge that gap is a fantastic application.
The one area I’d build upon is the idea of a "finite set of analysis." While the types of questions might be finite, the paths to the answers are not. As you mentioned, diagnosing a metric change can require navigating a huge, complex web of potential causes.
I think the core challenge here isn't a limitation of AI's raw intelligence, but rather our current inability as humans to effectively provide the AI with sufficient, precise, and constantly up-to-date context. Feeding an agent real-time knowledge of the system's architecture, the evolving business environment, recent feature launches, and external market factors is an immense engineering problem in and of itself. We don't yet know the best way to do it.
That being said, your proposed solution is a massive step in the right direction. It frames the problem perfectly, and focusing on storytelling is the right place to start. Thanks for putting this out there—it's a much-needed conversation.