Skip to content
Last9
Book demo

Analyze dashboard data with AI

Start AI analysis from a selected chart point or table row, review its context, and continue a focused investigation.

Use Ask Last9 AI on a dashboard chart point or table row to investigate a value in its service context. For example, start from a latency spike or an unexpected error count instead of typing the selected value and labels into a new conversation.

The action analyzes the selected point. To create or update a saved dashboard, see the separate dashboard creation workflow.

Start from a chart or table

  1. Open the dashboard and use its variables and time controls to locate the data you want to investigate.
  2. Click a data point in a supported chart or a row in a supported table.
  3. Select Ask Last9 AI from the point’s menu.
  4. Open the AI Analysis tab if it is not already selected. The analysis starts automatically.
  5. Read the response in the panel.

The action requires AI access and a point whose service context Last9 can identify. It may be absent for panels without a supported service or APM context. Markdown panels do not provide this action.

Understand the selected context

The analysis starts with information from the clicked point:

ContextWhat it contributes
Resolved metric queryThe selected series query with dashboard variables substituted.
Timestamp and valueThe observation you selected.
Metric labelsThe dimensions attached to that observation.
Service and environment, when identifiedLast9 identifies the service and environment from the available context.

Check that the service and environment in the result match the point you meant to investigate.

The dashboard time picker helps you find the observation, but the full dashboard time range is not passed as the investigation window. The analysis also excludes the dashboard’s other panels and its full variable configuration. If your question depends on additional dashboard details, provide them in a separate AI Assistant conversation.

Review the analysis

The analysis explains what the metric measures and the selected value. It assesses whether the value appears normal or concerning when the context supports that judgment, then describes possible causes and next investigation steps.

For example, a high latency value can be the starting point for checking slow requests. To determine whether latency increased over a period, compare time windows using the same service, route, and request filters.

The panel does not edit the chart, query, or saved dashboard, and you cannot type follow-up questions in it. The analysis is your own and has no share link. To give a colleague the findings, continue in AI Assistant and share that conversation.

Continue the investigation

For a follow-up, open AI Assistant and provide the relevant service, environment, timestamp, labels, and question. Include the chart query if you want to check its calculation:

I selected a latency spike for checkout-api in production at
14:12 UTC on 2026-09-10. Investigate 14:00 to 14:30 UTC and compare
it with the previous 30 minutes. Check slow traces and related
logs, and distinguish confirmed findings from possible causes.

Replace the example details with the point you selected. See Investigate an incident with AI for the follow-up workflow.


Troubleshooting

  • Ask Last9 AI does not appear in the point’s menu

    Ask an organization admin to check Enable Last9 AI in AI Settings. The action is available on supported charts and tables only.

  • The analysis has no service context

    Open AI Assistant and supply the service, environment, timestamp, and question directly.

  • The analysis covers the wrong data point

    Use the dashboard variables and time controls to locate the value you want, then select the point again. See Using Dashboards for variable and time selection.

  • You cannot ask a follow-up question in the panel

    Open AI Assistant and include the service, timestamp, labels, and the chart query.

Please get in touch with us on Discord or Email if you have any questions.