Beyond Dashboards: How to Build Living Data Experiences with Conversational Analytics
In my previous article, Why Most Business Dashboards Fail, I argued that dashboards — as we traditionally think of them — rarely deliver on their promise. They look great in a launch meeting, they generate some initial excitement… and then usage falls off a cliff.
The problem isn’t just bad dashboard design.
It’s that the whole idea of dashboards as the main way people interact with data is outdated.
See it in action: I built a working prototype that demonstrates these concepts.
The Dashboard Era Is Over
Dashboards were built for a world where:
- Data moved slowly
- Analysts were gatekeepers
- Business users were expected to “go into the tool”
- Insights were something you pulled when you had time
That world doesn’t exist anymore.
This article is the natural follow-up — a look at what comes after dashboards.
Welcome to the era of living data experiences, powered by proactive insights and conversational analytics.
The Core Problem: Dashboards Aren’t Where Decisions Are Made
Go through any company — manufacturing, retail, finance, tech — and you’ll find the same pattern:
- Someone requests a dashboard
- A BI team builds it
- It’s shown at a meeting
- Everyone nods
- And then… silence
High engagement, daily usage
Usage drops to near zero
Weeks pass. Usage drops to zero. New dashboards get requested. The cycle repeats.
Why Dashboards Fail
Dashboards fail not because people don’t care about data, but because:
- They require users to go somewhere else
- They require time, attention and navigation skills
- They often answer yesterday’s questions
- They rarely spark real action in the moment
In other words: Dashboards are good at presenting information, but terrible at delivering it.
Studies show that a staggering proportion of BI dashboards and analytics initiatives fail to gain traction or deliver sustained business value. Research indicates that up to 90% of Power BI dashboards fail to deliver meaningful value, while other sources note that “a staggering number of BI dashboards fail to gain long-term adoption.”
According to Gartner, Inc., new analytics paradigms (“agentic analytics”) are on the ascent — signalling that the dashboard-centric model is being superseded.
And that’s the real shift happening right now:
↓
To "getting people the insight they need at the moment they need it"
To get there, we need three major changes in how we think about analytics.
1. From Pull → Push: Data Should Come to You
Old Model: Pull
Most analytics today relies on a pull model:
- You log into a dashboard
- You navigate filters
- You try to interpret the charts
- You hope you're seeing something important
This is the digital version of "checking the weather" by going outside.
New Model: Push
Proactive insights delivered directly to the user.
Instead of checking dashboards for important shifts, the system notifies you:
"Sales for Account X dropped 18% vs last week.
Main driver: reduced orders in Product Line C.
Click to explore."
Or:
"Machine Z is experiencing a rising defect rate.
Probability of failure in next 7 days: 72%."
This reduces:
- Time to awareness
- Time to action
- Risk of missing critical changes
Dashboards become the context layer — not the entry point.
The entry point becomes: the insight itself.
2. From Visuals → Conversation: Talk to Your Data
Dashboards assume users know:
- Where to click
- Which filter to apply
- How to interpret the chosen visualization
This excludes all but the most confident power users.
Conversational Analytics Changes Everything
Conversational analytics turns that complexity into simplicity:
You just ask.
"Which products have rising returns but stable sales?"
"Show me resellers down more than 20% YoY."
"Drill into the Nordics."
"Show only robotic mowers."
"Highlight anything unusual in the last 14 days."
No training. No BI terminology. No dashboard fatigue.
Just you and the question you want answered.
Enter Databricks AI/BI Genie
Genie acts like an always-on analyst sitting on top of your Lakehouse:
- You type your question
- It understands what you mean
- It generates the SQL
- It runs the query
- It returns the answer with charts, text explanations and follow-up options
Every answer becomes a stepping stone for the next question.
It feels like having a smart analyst in every meeting.
And it changes everything.
3. From Projects → Products: Stop Shipping Dashboards, Start Shipping Experiences
This is the shift most organizations still need to make.
|
Dashboards Are Usually Shipped Like Projects
Result: 200 dashboards, 5% usage |
Living Data Experiences Are Managed Like Products
Result: Curated experiences, 80%+ usage |
Instead of building 200 dashboards for every possible use case, you focus on:
- A strong, governed data foundation
- A clean semantic layer
- High-value alerts
- Conversational access to everything else
- A small, curated set of dashboards for rituals (monthly business updates, board reports)
The goal is not to have “a dashboard for everything”.
The goal is to create the shortest path from question → insight → decision.
A Before/After Example
Before: Dashboard-Centric Analytics
A traditional sales dashboard:
- 24 charts
- 12 filters
- Information overload
- Used heavily the first month
- Practically abandoned by month 6
Managers end up asking for exports over email instead.
After: Living Data Experience
Automatic alerts for:
- Account drops
- Price anomalies
- Forecast deviations
- Spike in returns
Conversational analytics via Genie
Simple, focused dashboard for leadership meetings only
The New Workflow
- You receive an alert
- You click → Genie opens
- You ask follow-up questions
- You understand what happened
- You act immediately
No waiting for BI. No digging for dashboards. No delay.
Why This Is the Natural Sequel to “Why Most Dashboards Fail”
My previous article explains why dashboards fail:
- Wrong format
- Wrong expectations
- Wrong level of abstraction
This article explains what comes next:
- Push-based insights
- Conversational access
- Product thinking in analytics
- The end of dashboard sprawl
- Data that flows into decisions, not into PowerPoint decks
If the first post was about the symptoms, this one is about the cure.
Implementation Checklist
1. Audit your dashboard landscape
Kill dashboards nobody uses. Keep only the ones tied to business rituals.
2. Build a minimal semantic layer
Clear definitions. One source of truth.
3. Pilot conversational analytics
Start with one team. Sales, operations, supply chain — doesn’t matter.
4. Add proactive insight delivery
Threshold-based alerts → AI anomaly detection → real-time insight streams.
5. Shift BI to product thinking
Owner → roadmap → adoption metrics → iteration.
6. Measure time-to-insight
Your single most important KPI in modern analytics.
Final Thoughts: Dashboards Aren’t Dead — But They’re No Longer the Hero
Dashboards still matter. But they’re not the center of the data universe anymore.
The center is:
The real future of analytics is not “better dashboards”.
It’s living data experiences that:
- Push insights to you
- Let you talk to your data
- Fit naturally into your workflow
- Help you act in the moment
Dashboards explained the past.
Conversational analytics shapes the future.
And the gap between those two approaches is where the next generation of data-driven companies will win.
Related: Read the first part of this series: Why Most Business Dashboards Fail