Overview
An agent-assisted development course for experienced Power BI report creators. Attendees make controlled model and report changes, manage them with Git, and test and release Power BI content. Using a supplied starter, they also build and validate a simple read-only Fabric App connected to the same semantic model. Preview availability determines app-build access.
What you’ll need
- Windows laptop with current Power BI Desktop, headphones and microphone and reliable internet; Microsoft Edge or Google Chrome, Visual Studio Code and Git.
- Approved AI coding agent, suitable licence and usage allowance; course Power BI authoring tools and agent skills configured before attendance.
- Work account with required Power BI licence, Fabric capacity workspaces, and publishing, Git integration and deployment permissions.
- Course GitHub or Azure DevOps repository and separate development and test workspaces.
- Trainer-specified Node.js/npm, supplied starter, supported Fabric Apps preview environment and semantic-model/query API permissions (subject to preview availability).
Learning outcomes
By the end of the course, attendees will be able to:
- Translate a reporting requirement into an agent brief with acceptance criteria.
- Modify a semantic model and report with an agent; verify calculations, relationships and visuals.
- Version a Power BI project in Git; inspect changes, commit and restore a file version.
- Synchronise supported Power BI items with Fabric Git integration and explain its limits.
- Release a report and model from development to test; check connections, permissions and results.
- Build a simple read-only Fabric App with an agent and supplied starter, connected to a Power BI semantic model.
- Validate app figures against Power BI; test interactions, access, loading and error handling.
- Review and commit app changes, demonstrate the working app and explain deployment.
Course topics
Day 1 — Git and agent-assisted Power BI development
From business brief to controlled changes
- Define audience, business questions, measures and acceptance criteria; give the agent context, inspect its plan and set file, data and action boundaries.
- Introduce PBIP projects, PBIR report definitions and TMDL semantic models; initialise the supplied Git repository and save a baseline.
Developing and validating the model and report
- Inspect model metadata and create or refine DAX measures with Power BI authoring tools; use report-authoring agent skills to adapt pages, visuals and formatting.
- Check totals, filter context and relationships against known results; spot invented fields and incorrect calculations. Validate layout, interactions and accessibility in Power BI Desktop.
- Review Git diffs, commit accepted changes and restore a file version; introduce branches and pull requests. Keep sensitive data and credentials out of repositories and agent conversations.
Day 2 — Fabric DevOps
Testing and releasing Power BI content
- Connect a prepared workspace to Git and synchronise supported report/model definitions; distinguish these from data, credentials, permissions and environment settings.
- Introduce development, test and production stages, dependencies and approval; deploy the course report and model to test with a Fabric deployment pipeline.
- Check connections, refresh, measures, report behaviour and a supplied security role; record release evidence and recovery plans. See automated validation without building enterprise CI/CD.
Day 3 — Building a simple Fabric App with agents
- Distinguish custom Fabric Apps from Power BI report-distribution apps; introduce Rayfin, Microsoft’s Fabric Apps development framework.
- Use the supplied Rayfin starter and authenticated connector to connect to the course semantic model; identify interface, data access and hosting.
- Guide an agent to build a read-only interface with KPI cards, a chart, data grid and filtering; inspect code and commit reviewed changes.
- Run and demonstrate the working app; compare figures with Power BI and test filtering, sign-in, loading, empty and error states.
- Observe deployment with the Rayfin CLI, distinct from pipelines for supported Power BI items.
Note: This new, rapidly evolving course is likely to be updated as AI capabilities, Power BI and Fabric Apps develop.