Back to Cube

Dashboard Agent

docs-mintlify/docs/explore-analyze/dashboards/dashboard-agent.mdx

1.6.567.2 KB
Original Source

The Dashboard Agent is a conversational assistant docked as a panel on a published or embedded dashboard. Ask it questions in plain language about what's on the dashboard — and it answers using live queries against your semantic model, not just the numbers already rendered on screen.

Because it queries the model directly, the agent can go beyond what a chart shows. It can drill into a single chart, re-aggregate a metric, break a number down by a dimension, pull a longer time range than the chart displays, and summarize trends across the dashboard.

It can also apply filters and change the time granularity on the dashboard you're viewing — just ask in plain language (for example, "filter to last 30 days" or "group by month"). These changes affect only your own session and are never saved to the dashboard — see Changing filters and time granularity.

<Frame> </Frame> <Note> There are **two different agent surfaces** with different capabilities:
  • Dashboard Agent (this page) — for viewers of a published or embedded dashboard: question-and-answer about what's on screen, plus changing your own filters and time granularity.
  • Workbook Agent — full authoring inside a workbook: creating reports, building and editing dashboards, and running analysis.

The published Dashboard Agent can adjust your own view — filters and time granularity, for your session only — but intentionally cannot author or change the saved dashboard. </Note>

Opening the agent

Open a published dashboard and toggle the agent panel from the dashboard header to start a conversation. Use New Chat to start a fresh thread.

The panel is docked to the right side of the dashboard and is resizable — drag its edge to give the conversation more or less room. On narrow screens it switches to a compact mode so it stays usable next to the dashboard.

Focusing on a specific chart

<Note> Chart-level follow-ups and attached-chart chips are **rolling out**. If you don't see the **Ask a follow up question** action on a chart's menu yet, it hasn't reached your deployment. </Note>

To ask about one chart in particular, open that chart tile's menu () and choose Ask a follow up question. This opens the agent with that chart attached as context, so your question is scoped to that chart's data.

Attached charts appear as removable chips in the chat input. To drop a chart's focus, remove its chip. You can attach more than one chart to ask about them together.

When a chart is attached, the agent receives a read-only snapshot of it — the underlying SQL, the data, and the chart spec — along with the dashboard's current filter and time-granularity state. It uses this as context for answering and never edits the chart's definition. You can still ask it to change the dashboard's filters or time granularity — see Changing filters and time granularity.

Changing filters and time granularity

Ask the agent, in plain language, to filter the dashboard or change a time granularity, and it applies the change to the controls in front of you:

  • "Filter this dashboard to orders with status completed."
  • "Show only the Outerwear category."
  • "Filter to the last 30 days."
  • "Group by month."
  • "Reset the filters."

How it works and what to expect:

  • Only controls that exist on the dashboard can be changed. The agent can change a filter or time granularity only when the dashboard has a matching filter or time-granularity control widget for that dimension. If there's no control for what you asked about, the agent tells you so instead of changing anything. A control whose visibility is Disabled also can't be changed.
  • Time granularity stays within the control's allowed granularities. If a time-granularity control restricts which granularities viewers can pick, the agent only chooses from that set.
  • Changes apply to your session only. They're applied the same way as changing a control by hand: reflected in the dashboard's URL — so you can share or bookmark the filtered view — but not saved to the dashboard. They don't affect what other viewers see, and the author's saved defaults are left untouched.

This works the same way on both published dashboards and embedded dashboards.

Attachments

You can attach uploaded files to the conversation, and the agent can read their contents — text, PDF, ZIP, and image files are supported.

<Note> Pasting an image directly from your clipboard into the chat may also be supported. Confirm this in your deployment before relying on it. </Note>

What the Dashboard Agent can and can't do

The published Dashboard Agent can answer questions and adjust your own view, but it never changes the saved dashboard.

It can:

  • Answer questions about what's on the dashboard from the dashboard's own data
  • Run read-only queries against the semantic model to go deeper — drill down, break a metric down by a dimension, or pull a longer time range than a chart shows
  • Apply filters and change the time granularity on the dashboard you're viewing, for dimensions that have a matching control widget (your session only — see Changing filters and time granularity)
  • Explain how a metric or dimension is defined
  • Look up the values available for a dimension
  • Read the contents of attachments you add to the chat

It can't:

  • Save its filter or time-granularity changes to the dashboard — they apply to your session only and never change what other viewers see
  • Edit the dashboard, or add/remove widgets
  • Create reports or workbooks
  • Change the data model

If you need any of those, use the Workbook Agent instead.

Limitations

  • Filter and time-granularity changes need a matching control. The agent can change a filter or time granularity only for dimensions that have a corresponding control widget on the dashboard, and the change applies to your session only — it is never saved to the dashboard. If you ask it to change something with no control on the dashboard, it explains the current state instead of applying a change.
  • No authoring on published dashboards. The published Dashboard Agent cannot create reports or build dashboards, edit the saved dashboard, or change the data model. Those live in the Workbook Agent.
  • Report search covers published reports only. When the agent searches for existing reports, it sees published reports.