Creating a dashboard step by step

Before you start

You need at least one connected data source. See uploading files or connecting a database.

Three ways to get a dashboard

1. The automatic insights dashboard

When a data source finishes its first sync, clariBI builds a dashboard for it: headline numbers and charts for the columns it found. You do not have to do anything. Open AI Dashboards in the sidebar to see it. This is also the dashboard the Free plan keeps.

2. Create a dashboard with AI

Open AI Dashboards and click Create Dashboard. Enter a Dashboard Name, an optional description, and pick a Dashboard Type:

Dashboard Builder (recommended)

A five-step builder:

1. Data Sources: select one or more connected sources.
2. Business Domain: pick the industry or area, such as e-commerce or SaaS. clariBI uses it to suggest metrics.
3. Define Metrics: choose from Suggested Metrics or Quick Start Templates, or click Create Custom Metric and describe it in plain language (for example "Revenue per order by month"). For each metric pick a visualization type.
4. Design Layout: place the widgets on the grid.
5. Deploy: pick a Refresh Frequency (manual, hourly, daily, weekly or monthly), optionally share with workspace members, and check the estimated credits per refresh before you create the dashboard.

Quick AI-generated dashboard

Describe what you want under Data Context (for example "Sales performance with revenue trends and top customers"), select the data sources and create it. clariBI picks up to 15 widgets and writes a short narrative for each. Every number in the narrative is checked against the chart data.

Both options run as a background job; you can keep working and are notified when it is ready.

3. Manual mode

Manual mode builds a dashboard from a dashboard template without AI, so it uses no credits. On the Trial and paid AI plans it is the Manual mode button next to Create Dashboard. On Lite, Create Dashboard opens it directly.

1. Pick a template: search the dashboard templates or filter by category.
2. Pick data sources: select the sources to feed it.
3. Bind columns: for each slot in the template, choose the column that fills it (for example which column is the date and which is revenue). A preview shows the result. If a binding does not fit, the wizard lists what to fix.
4. Click Create Dashboard.

After creating

On the dashboard page you can:

• Refresh the data.
• Open Settings to rename it, change the refresh frequency, see the estimated monthly credit use and change the widget layout.
• Share Link to manage sharing (public links need Professional or Enterprise).
• Share to Workspace (Professional and Enterprise).
• Export the dashboard's numbers as CSV (one row per value) or JSON (grouped by widget) on Professional and Enterprise.
• Forecast Dashboard to forecast every widget that can be forecast (Starter and above).
• Track a widget's metric as a goal, or link a widget to an existing goal.
• Comment in the discussion panel when the dashboard is in a workspace.

See sharing dashboards and dashboard layout.

Tips

1. Start from the question. Decide what the dashboard should answer, then choose metrics for that.
2. Keep it focused. Five to eight widgets are easier to read than fifteen.
3. Put the main number first, top left.
4. Use one date range across widgets you want to compare.
5. Pick the refresh frequency you need. Each AI refresh uses credits; daily is enough for most dashboards.

The documentation's build your first dashboard is a shorter version of these steps, and build a sales dashboard from scratch is a worked example.

Related articles

• Dashboard builder reference
• Widget types
• Dashboard templates