Fabric data Agent visualization capability and learning capability

[ann33]

Hi 

I have two question 

1.Is the Fabric Data Agent designed to learn and get better with usage?

2. Are there any plans to include visualization capability in Data agent simiar to databricks genie?

   Is there any way we can achieve the visualization capability ?

[v-pnaroju-msft]

Hi ann33,

We are following up to inquire whether you have submitted this as an idea in the Ideas Forum. If so, kindly share the link here, as it would be beneficial to other community members who may have similar feedback.

Should you have any further queries, please feel free to contact the Microsoft Fabric community.

Thank you.

[v-pnaroju-msft]

Thank you, @rohit1991 and @Vinodh247  , for your response.

Hi ann33,

We sincerely appreciate your inquiry made through the Microsoft Fabric Community Forum.

Based on my understanding, the Fabric Data Agent is currently designed primarily for querying and orchestrating data across various sources and does not natively support built-in visualizations or custom UI extensions. The current functionality of the Data Agent focuses on secure data access and automation, rather than rendering visuals or embedding into custom interfaces. Microsoft Fabric does provide advanced visualization options through Power BI and Notebooks; however, these are separate components distinct from the Data Agent.

As this is a valuable enhancement suggestion, we encourage you to kindly submit it on the Microsoft Fabric Ideas Forum via the link Fabric Ideas - Microsoft Fabric Community so that the product team may evaluate it for inclusion in future roadmap planning.

Additionally, please find below some useful links for your reference:
Fabric data agent Python SDK (preview) - Microsoft Fabric | Microsoft Learn

Fabric data agent scenario (preview) - Microsoft Fabric | Microsoft Learn

We hope that the information provided will assist in addressing your query. Should you require any further assistance, please do not hesitate to reach out to the Microsoft Fabric community.

Thank you.

[ann33]

can we build a custom UI , which will call the data agent SDK on top of fabric? If yes, what could be the suggested flow.

[Vinodh247]

1. Is the Fabric Data Agent designed to learn and get better with usage?

• No, the Fabric Data Agent does not currently learn or improve through usage in the sense of reinforcement learning or fine tuning.

• It is prompt-driven, powered by Azure OpenAI (typically GPT models) and does not retain memory between interactions.

2. Are there any plans to include visualization capability in Data Agent similar to Databricks Genie?

as of now:

• Microsoft has not yet announced built-in visual output features within the Fabric Data Agent itself.

• Databricks Genie can generate charts inline in notebooks using code generation Fabric's Copilot in Notebooks supports this, but Data Agent does not yet have visual rendering built-in.

Please 'Kudos' and 'Accept as Solution' if this answered your query.

[rohit1991]

Hi @ann33 ,

1. Learning Capability: Currently, the Fabric Data Agent does not have built-in machine learning or adaptive features that help it “learn” and improve based on usage over time. It mainly connects data sources and manages pipelines but doesn’t self-optimize or learn from patterns like an AI assistant.
   
   

2. Visualization Capability: At the moment, the Fabric Data Agent doesn’t offer native visualization features similar to Databricks Genie. Visualization typically needs to be handled in Power BI, Fabric notebooks, or other reporting tools connected to your Fabric workspace.

   
   Workarounds:

   • You can output results from Data Agent pipelines to a dataset and visualize them in Power BI.

   • For interactive/AI-powered visualizations, you can use Microsoft Copilot features in Power BI (if available) or leverage Python/R scripts in Fabric notebooks.
     
     

As of now, there’s no official announcement about adding direct visualization to Data Agent itself, but Microsoft is actively updating Fabric, so this could change in the future.