Natural Language Chat Agent¶
This page introduces the chat agent (Dashboard Builder) feature in KX Dashboards, explaining what it does and how it can help you build and edit dashboards faster.
The Dashboard Builder lets you create and edit dashboards using plain-language descriptions rather than manual configuration. Instead of dragging components onto a canvas and wiring up data sources by hand, you describe the dashboard you want and the chat agent generates it for you automatically.
This is powered by your organization's configured AI provider (OpenAI, Claude, or Gemini). When you submit a request, KX Dashboards sends the AI provider the context it needs to produce a valid result — including details about available data sources, table metadata, column descriptions, and the rules for constructing a well-formed KX dashboard. The provider returns a JSON dashboard definition, which is loaded directly into the application.
Key capabilities¶
- Create dashboards from a description — Describe the visualizations, data, and layout you want in natural language. The agent generates a working dashboard based on your query.
- Edit an existing dashboard — Describe the changes you need to make to an open dashboard. The agent modifies it without requiring you to locate and adjust individual component properties.
- Specify a target table — Optionally name the table you want to query. This lets the agent scope its generated queries to a specific data source.
- Browse chat history — Your sessions are saved automatically. You can reopen previously generated dashboards directly from the Chat Agent dialog via clickable links.
- Delete chat history — Permanently remove your session history when it is no longer needed.
Benefits¶
Speed — Producing a functional multi-component dashboard that would otherwise take several minutes of configuration can be done in a single natural language prompt.
Accessibility — Users who are unfamiliar with the full range of dashboard components, query syntax, or data source configuration can still build useful dashboards without deep product knowledge.
Iteration — Because you can edit an open dashboard by describing the changes you want, it is easy to refine a layout or adjust a visualization without having to rebuild from scratch.
Data awareness — The agent has access to your data source metadata and column descriptions, so it can generate queries that are grounded in your actual data schema rather than generic placeholders.
How it works¶
The diagram below summarizes the flow from a natural language prompt to a rendered dashboard.
- You type a description in the Dashboard Builder dialog.
- KX Dashboards assembles a prompt containing your request, the data source skill (datasource metadata and column descriptions), and the rules for generating a valid dashboard definition.
- The assembled prompt is sent to your configured AI provider.
- The provider returns a JSON dashboard definition.
- KX Dashboards loads the definition, and the dashboard opens in the application.
Requirements¶
An administrator must configure the chat agent before it can be used. Configuration is done via Global Properties → NLX tab and requires an AI provider, a model, and an API key. See Configure the Chat Agent for full setup and usage instructions.