Agentic development is a new paradigm where developers shift from writing code or using UI applications to guiding intelligent agents. The developer defines the what - the requirements, rules, and intent - and the AI agent handles the how - generating specs, implementing the model, running validations, fixing issues, and iterating toward a working solution.
You can achieve agentic development in Power BI combining the following features:
- Open file formats with Power BI Project file format (PBIP) with TMDL language and Power BI enhanced Report format (PBIR).
- MCP Server powerbi-modeling-mcp
- Ensure the AI has context information on your Power BI development style. See .kb/ for an example.
Important
powerbi-modeling-mcp is currently in private preview. If you're interested in trying it out, sign up here.
See the branch final to inspect the final output.
- Create a high-level requirement docs, similar to requirements.md.
- Prompt AI agent to create a new semantic model but create a spec before implementation. prompt example. Give it context about your Power BI development style.
- AI creates a development spec – readable, reviewable, and adjustable.
- You review the spec, and when OK. Just say GO and the agent starts implementing the spec autonomously.
- You can ground the development with Best Pratice Analysis (see .bpa/). And autonomously fixes issues and iterated until there are no issues.
- You open the Power BI application and review the result, just like you would with a teammate’s pull request.
Important
Its very important to provide AI knowledge/context about your Power BI development style and Power BI concepts such as PBIP structure. See .kb/ folder.
- Clone this Repo into your laptop
- Install Visual Studio Code
- InstallGitHub Copilot and GitHub Copilot chat
- Open Github Copilot Chat in Agent mode
- Copy and paste the prompt into the chat.
- Review the generated development spec, then prompt Copilot to proceed with implementation.
