Dashboards transform raw data into dynamic, interactive visuals that enable teams to track KPIs, monitor performance, and support decision-making at scale.

While often seen as the final step in an analytics workflow, effective dashboards are best understood as living interfaces—not static reports. When built well, they surface the most relevant insights at the right level of detail, tailored to specific users or decisions.

TrueState makes it easy to design, test, and deploy AI-assisted dashboards that integrate natural language exploration, automated commentary, and chart generation. This guide walks you through how to think about, structure, and iterate on high-impact dashboards using the platform.

​
What makes a great dashboard?

Dashboards are most valuable when they are:

  • Purpose-driven – designed around specific decisions or stakeholder goals
  • Minimalist – showing only the most relevant information
  • Interactive – allowing users to filter, segment, and drill down into areas of interest
  • Up-to-date – refreshed regularly and responsive to underlying data changes
  • Contextualised – paired with clear commentary and business framing

Before building, ask: Who will use this dashboard? What questions should it help them answer in under 30 seconds?

​
Dashboard use cases

There are three common use cases for dashboards in TrueState:

​
1. Operational monitoring

Track critical metrics that need daily or weekly oversight—like revenue pacing, lead volume, churn risk, or support backlog.

Tips:

  • Use time-series visuals (e.g., line or area charts) to show recent performance vs targets.
  • Add threshold indicators (e.g., conditional formatting or alerts).
  • Keep it lean—focus on action-triggering metrics, not vanity data.

​
2. Strategic performance reviews

Support monthly or quarterly reviews with charts that compare teams, cohorts, or markets over time.

Tips:

  • Include breakdowns by department, geography, or customer segment.
  • Use bar and heatmap visuals for comparisons.
  • Embed AI-generated summaries that auto-interpret trends and changes.

​
3. Exploratory deep-dives

Empower teams to explore trends in self-serve mode without needing to write queries.

Tips:

  • Use filters, slicers, and dropdowns to enable exploration.
  • Include freeform search or natural language prompts for open-ended questions.
  • Keep data sources well-labelled to ensure usability.

Dashboards aren’t just visual—they’re interfaces for curiosity. Consider how different personas will navigate and use them.

​
Building dashboards with agents

The fastest way to build dashboards in TrueState is by collaborating with an AI agent directly in the dashboard builder.

​
How it works:

When you’re in the dashboard creation interface, you’ll see a “Dashboard Assistant” on the left-hand side. You can interact with it using natural language prompts.

For example:

  • “Show total sales by region over the last 6 months.”
  • “Create a funnel for lead conversion across marketing stages.”
  • “Add a chart comparing churn rate for Q1 vs Q2 by product tier.”

The agent will respond with one or more suggested charts and commentary, ready to be inserted directly into your layout.

You can then:

  • Edit the chart config (dimensions, filters, titles)
  • Convert the visual into a reusable component
  • Group charts into tabs or sections
  • Save as a template for other dashboards

Use the agent to prototype layouts quickly, then refine with manual tweaks. It helps reduce the blank canvas problem and jump-start your design.

​
Live interrogation with the Dashboarding agent

Once a dashboard is live, TrueState enables live interrogation of both the dashboard elements and the underlying dataset using the Dashboarding agent.

​
Use cases:

  • Get a summary of trends across multiple charts

    “What changed in revenue performance last month?”

  • Ask follow-up questions based on visible patterns

    “Why did churn spike in March for enterprise customers?”

  • Compare dimensions on the fly

    “Compare win rate between paid and organic leads.”

The agent understands both the dashboard structure and the underlying dataset, allowing for real-time responses that reference current filters, context, and prior queries.

Think of this as having a virtual analyst embedded in every dashboard. You can ask questions without needing to leave the interface or write SQL.

​
Agent capabilities:

  • Cross-chart synthesis (e.g., “Summarise the key drivers of Q2 performance.”)
  • Automated anomaly detection (e.g., “Any metrics outside expected bounds?”)
  • On-demand calculations or comparisons
  • Fast “why” and “what-if” scenario generation

You can also pin agent responses to the dashboard for others to view—this is helpful for shared understanding in team reviews.

​
Workflow for building dashboards in TrueState

  1. Define the goal
    What decisions or actions should the dashboard support?

  2. Identify key metrics
    Choose 3–5 KPIs. Avoid the temptation to include everything.

  3. Select and prepare data sources
    Ensure datasets are clean, up-to-date, and well-labelled.

  4. Use the Dashboard Assistant
    Prompt the agent for initial chart ideas and layout suggestions.

  5. Refine and annotate
    Add or edit visual elements manually. Use commentary and notes.

  6. Test with users
    Share draft dashboards, get feedback, and validate filter logic.

  7. Launch and interrogate
    Publish the dashboard and explore it with the Dashboarding agent.

  8. Iterate over time
    Monitor usage, update metrics, and adjust based on business shifts.

​
Best practices

  • Limit cognitive load: Avoid more than 6–8 charts per view.
  • Group related metrics: Use tabs or sections to separate domains.
  • Label everything: Axes, filters, and legends must be unambiguous.
  • Tell users what matters: Use agent commentary or manual notes.
  • Make time dynamic: Always include a time selector or refresh period.
  • Support mobile: Design for responsiveness if dashboards are used on mobile.

​
Glossary

  • Dashboard Assistant – The AI agent that helps generate charts and layouts during dashboard creation.
  • Dashboarding agent – The AI agent used to interrogate live dashboards and underlying datasets.
  • KPI (Key Performance Indicator) – A quantifiable measure used to track progress toward a goal.
  • Annotation – Commentary embedded in a dashboard to explain changes, trends, or anomalies.
  • Drill-down – The ability to explore data at more granular levels (e.g., from company to department to employee).
  • Vanity metric – A number that looks good but doesn’t inform a decision or action.

​
Common pitfalls to avoid

​
1. Overloading with metrics

Users get overwhelmed. Focus on a few key indicators that matter.

​
2. Ignoring user intent

Build dashboards for specific audiences. A CFO and a product lead need different data.

​
3. Underutilising agents

The dashboard experience is best when you treat agents as collaborators. Don’t ignore them.

​
4. Skipping QA

Always test filters, refresh logic, and labels before publishing. Broken dashboards erode trust.

​
Next steps

  • Open a new dashboard in TrueState and activate the Dashboard Assistant
  • Ask for a layout based on your current KPIs and business priorities
  • Use the Dashboarding agent to explore your charts and uncover insights
  • Publish to your team and annotate key takeaways