FRENES
Book a call

Global luxury group

A Git-first CI/CD framework on GitHub Actions, now the standard way the whole BI team releases.

Several teams of Power BI developers, manual releases, and no systematic check before production.

Luxury
  • Power BI
  • Microsoft Fabric
  • GitHub Actions
  • CI/CD

Key results

of Power BI developers trained
100%
environments, DEV to PRD, with controlled promotion
3
central configuration file instead of per-environment secrets
1

Context

Several teams of Power BI developers, manual releases, and no systematic check before production.

The challenges

  • Manual releases, done differently by each developer team.
  • No systematic check before a change reached production.
  • Secrets and settings duplicated in every environment.
  • Variants of the same model maintained by hand.

What we did

Industrializing Power BI and Fabric for a global luxury group

  1. A Git-first CI/CD framework

    Models and reports versioned in TMDL and PBIR, GitHub Actions pipelines authenticated with Azure OIDC, controlled promotion from DEV to UAT to PRD.

  2. Quality gates on every change

    Model validation and best-practice rules through Tabular Editor over XMLA, deployment risk controls, DAX smoke tests and targeted refresh validation.

  3. Workspaces provisioned automatically

    Capacity, access and Git integration configured through the Fabric REST APIs: no more hand-built workspaces.

  4. One master model, synchronized variants

    Changes to the master model flow to its variants. One central configuration file replaces per-environment secrets.

  5. Self-sufficient teams

    Every Power BI developer trained on the new way of releasing, standards documented.

Architecture

From commit to production, with no manual stepEvery change follows the same path: versioned in Git, checked automatically, then promoted from one environment to the next.
  1. Git
    • TMDL models
    • PBIR reports
  2. CI checks
    • Best practices
    • DAX tests
    • Deployment risks
  3. DEV
    • Provisioned workspace
  4. UAT
    • Refresh validation
  5. PRD
    • Controlled promotion
  • Central configuration
  • Azure OIDC
  • Fabric REST APIs

Before

  • Manual releases
  • Different practices in each team
  • Workspaces created by hand
  • Secrets duplicated per environment

After

  • Automated DEV → UAT → PRD promotion
  • One shared release standard
  • Workspaces provisioned through APIs
  • One central configuration

What changed

A release is no longer a risky event: every change passes the same checks, and the whole BI team releases the same way.

Technologies
  • Power BI
  • Microsoft Fabric
  • GitHub Actions
  • TMDL / PBIR
  • Tabular Editor
  • XMLA
  • Azure OIDC
  • Fabric REST APIs
  • PowerShell
Format
Industrialization engagement within the tech division, with several Power BI developer teams.

Contact

Are your Power BI releases still manual?

We set up the same delivery pipeline for you, adapted to your teams and your tooling.

Book a callExplore BI Factory[email protected]

Case studies

More case studies

  • Healthcare / Medtech

    Medtech group

    Moving beyond Excel: a medtech group's first BI platform

    Context
    A medtech group with no BI platform: sales, marketing, finance and leadership ran the business from large Excel files and exports reworked by hand.
    Outcome
    130+ users across sales, marketing, finance and leadership-50% manual data handling
    • Power BI
    • CRM
    • Self-service
    • Automation
  • Jewelry

    International jewelry house

    Running BI for an international jewelry house

    Context
    Reporting on SAP BusinessObjects and SQL Server to migrate to Power BI and Google Cloud for finance, omnichannel and supply, with thousands of users worldwide.
    Outcome
    4,000–5,000 self-service users worldwideAround 20 standardized report templates on certified data
    • Power BI
    • BigQuery
    • Microsoft Fabric
    • Migration