Dashboard AI assistant

Last updated 10/02/2026

Deploying AI capabilities

After you upgrade to v5.0, contact the service team to deploy the AI capabilities and turn on the AI features you need before you can use them.

Starting from v5.0, dashboards integrate AI capabilities that help you interpret data and get data insights faster.

Dashboard AI interpretation

Click the dashboard icon with a star in the dashboard title bar (its hover tooltip is Dashboard Analysis). A menu appears first, with a Please enter your thoughts input box and a ×× dashboard interpretation item. Interpretation starts only after you click ×× dashboard interpretation.

By intelligently analyzing dashboard data, AI interpretation provides:

• Data description: presents the trends, distributions, and correlations of key metrics in a structured way;
• Anomaly detection: automatically identifies data fluctuations and outliers;
• Value summary: extracts the business insights behind the data and gives actionable recommendations;
• Interactive exploration: lets you ask follow-up questions about the interpretation to dig deeper into the value of the data

You can turn the interpretation result into a note with Save as note, and the note is displayed at the top of the dashboard.

When you use it, keep the following in mind:

1. Supported data

   • ✅ Supported components: most model reports and tags
   • ❌ Unsupported components: Scenarios and Flows Analysis reports, and notes

2. Data loading: make sure the dashboard data is fully loaded

   • For a long dashboard, scroll through it to confirm that all reports are displayed properly;
   • If the data isn't fully loaded, the interpretation can only be based on the loaded data, which may lead to biased conclusions

3. Data volume limit

   • If the data volume exceeds what the large language model can process, part of the data is automatically sampled for interpretation;
   • In this case, the interpretation may have limitations, so use the results with caution.

Tip: When dashboard names and report names are accurate and convey information about the business scenario, the interpretation fits the context better

Report AI interpretation

For the data of a single report, AI interpretation analyzes the data returned by the report around the report's topic. You can ask follow-up questions about the results to better understand how the data performs. The entry is on the report card (its hover tooltip is Report interpretation). Clicking it also opens a menu first, and interpretation starts only after you click the interpretation item in the menu.

If you need more analysis dimensions, add group-by items to the report. The interpretation can't use data outside the report results.

Dashboard generation assistant

You can find the dashboard generation assistant on the left side of the dashboard. Enter an analysis topic you care about, such as "Analyze user activity", and the AI assistant breaks down the analysis approach and builds related reports for you with the analysis models of AE. In the conversation, you can ask the AI assistant to keep refining the dashboard or report content, for example, adding, editing, or deleting reports, adding metrics of interest, changing the analysis period, or renaming the dashboard.

Feature entry

Generate dashboard content from a question