What are Generative Dashboards?
Generative Dashboards allow users to create fully functional dashboards in Alli using a simple natural language prompt. Instead of manually setting up charts, users can describe what they want to see, and Alli will automatically build a dashboard using live cross-channel marketing data.
Who can use Generative Dashboards?
All users with access to Alli’s Dashboards application can use Generative Dashboards. Permissions are tied to Alli’s unified access system:
-
Client Strategists and Analysts can create, edit, and save dashboards.
-
Clients and view-only roles can access dashboards shared with them but cannot create or edit.
What data sources does it use?
All dashboards are powered by the Cross-Channel View in Alli, which combines data from platforms like Meta, Google, TikTok, Amazon Ads, and more. This ensures consistent, comparable metrics across channels.
If cross-channel has not been set up it defaults to the ad_performance Certified datasource
How can I validate the results?
-
Use the exportable data table to compare numbers against numbers you see in platform
-
Filters in Generative Dashboards apply to the entire dashboard, so ensure you’re comparing on the same dimensions (e.g., time range, channel).
How can I add a new or custom data source?
Currently, Generative Dashboards only support Alli’s Cross-Channel View. Custom or client-specific data sources (like feeds, CRM, or offline data) are not yet supported. if cross channel is not setup it defaults to Ad_performance certified datasource
Can I copy dashboard templates from other clients into my client?
Not today. Each client’s dashboards are unique to that client’s data. You can, however, duplicate a dashboard within the same client (via the three-dot menu → Duplicate) and edit filters to create variations like campaign-specific or channel-specific views.
What kinds of dashboards can I create?
You can create:
-
Campaign recaps (performance summaries)
-
Cross-channel comparisons
-
Executive summary dashboards
-
Channel-specific performance views (e.g., paid social, search, retail media)
How long does it take to generate a dashboard?
Most dashboards generate within a few minutes, depending on the complexity of the request and dataset size.
Can I edit the dashboards after they’re generated?
Yes. Users can:
-
Edit one visualization at a time (chart type, columns, colors, and titles)
-
Use the Code View to customize or fully rebuild a chart using JavaScript (with Highcharts support)
-
Apply filters at the dashboard level (filters apply to all visualizations, not individual charts)
Can I add new charts after creating a dashboard?
Not yet. Current functionality only allows editing existing visualizations. Adding new visualizations will be introduced in a future release.
Can I export dashboards?
You can export the data table for dashboards. Visualizations themselves cannot be exported.
What happens if the dashboard doesn’t look right?
-
Use the edit functionality to adjust chart types or metrics.
-
For deeper changes, switch to Code View.
-
If errors appear in Code View, you can copy them into an AI assistant (like Alli Assist Bar) or adjust the code directly.
How should I write prompts?
Use terms in Glossary
Best practices:
-
Be specific about the metrics (e.g., “Spend, impressions, CTR”)
-
Include the level of breakdown you want (e.g., “by channel and campaign”)
-
Mention the story you want the dashboard to tell (e.g., “Executive summary of last month’s media performance”)
What are the current limitations?
-
Cannot add new visualizations post-creation
-
Only supports Cross-Channel data source (no client-specific feeds yet)
-
Export is limited to the table, not visualizations
-
Dashboards cannot yet be duplicated across clients
Where can I get help?
https://help.alliplatform.com/dashboards/Working-version/how-to-get-help-with-gen-dashboards
-
Help Docs on Alli Hub (step-by-step guides and glossary)
-
Slack channel #alli-gen-dashboards for quick questions
-
Alli Assist Bar daily office hour (10:20–11:20 CST) for live support with engineers (bring your analyst if data-specific)