Create dashboards using AI
AI-powered tools let you create dashboards from natural language, without writing API requests or learning the dashboard schema. Use them to go from a question to a working dashboard through conversation, or to integrate dashboard generation into your own AI workflows.
| Tool | When to use this | What you get |
|---|---|---|
| Elastic Agent Builder | Creating dashboards through conversation in Kibana or using the MCP server | Dashboard through chat that you save when ready |
| Kibana dashboards agent skill | Building your own AI agent or LLM tool that generates dashboards | Saved dashboard (using the API) |
If you need to create dashboards from code or CI/CD pipelines using REST APIs, refer to Create dashboards programmatically instead.
Elastic Agent Builder agents can create and update dashboards through natural language chat, using the chat UI in Kibana, the Chat API, or the MCP server. Describe what you want to visualize and the agent builds a dashboard with ES|QL-powered visualizations. Dashboards are in-memory by default and exist as conversation attachments until you save them, so you can iterate freely before finalizing.
Use Elastic Agent Builder when you want to:
- Go from a question to a working dashboard without writing API requests or learning the schema
- Explore an unfamiliar data source by asking the agent to surface and visualize key fields
- Prototype a dashboard through conversation, then save it when you are satisfied
Elastic Agent Builder generates ES|QL-powered visualizations, markdown panels, and collapsible sections. For panel types the agent does not support yet, such as controls, use the Dashboards API directly.
Full-screen standalone chat mode gives the most space for previewing and refining dashboard content.
Refer to Chat with Elastic Agent Builder agents.
The kibana-dashboards agent skill is an open-source skill for integrating dashboard generation into your own AI tools. It provides language models with the context and instructions to generate valid dashboard definitions and call the Dashboards API, so you can build custom interfaces, automation scripts, or agentic pipelines that create Kibana dashboards without relying on the built-in Elastic Agent Builder experience.
Use the agent skill when you are:
- Building a custom AI application that creates dashboards as part of a larger workflow
- Integrating dashboard generation into an agentic pipeline outside the Kibana UI
- Extending how an LLM interacts with the Dashboards API beyond the built-in capabilities of Elastic Agent Builder
The Elastic Agent Builder built-in tools use a similar mechanism internally. The agent skill is for teams who need the same capability in their own tooling.
Pair it with the elasticsearch-esql skill to give the agent the ability to discover available indices and fields before generating ES|QL queries for dashboard panels.