Migration guide

    Leave the metric mess behind when you leave Power BI

    Moving off Power BI starts with an inventory of the reports and semantic models people still use. Rebuild those metrics in a warehouse semantic layer, run both in parallel, then cut over. A new tool full of copied reports still carries the old metric mess. Teams often use an AI-native data and systems integrator like Rig to run the inventory, mapping and cutover.

    By Toby Mather · Published

    Key takeaways

    • Power BI keeps its business logic in semantic models, as DAX measures, relationships and row-level security roles, so the models are what you migrate.
    • Pull usage from the Power BI activity log and workspace usage reports before you decide what to rebuild.
    • Rebuild the measures once in a semantic layer on the warehouse, so the next BI tool and any AI agents read the same definitions.
    • Tableau, Looker, Sigma, Omni and Apache Superset all appear as destinations, and the right one depends on who builds the reports.
    • If you are staying in Microsoft and moving to Fabric, the same inventory applies and the Azure Synapse guide covers the platform choice.
    1. 01Power BI reports and models

      DAX, relationships and RLS roles

    2. 02Inventory

      Usage, owners and measures

    3. 03Warehouse semantic layer

      One definition per metric

    4. 04Parallel run and cutover

      Totals and access compared

    5. 05Optional: move to Fabric

      If you are staying in Microsoft

    The Power BI estate moves to a warehouse semantic layer, with an optional Fabric path.

    How do we inventory reports, datasets and semantic models?

    The Power BI activity log and the usage metrics reports show which reports people open and how often. Combine that with a list of workspaces, semantic models and dataflows from the admin APIs, and you have the starting inventory. Shared semantic models matter most, because one model can feed dozens of reports.

    For each model that survives, copy out the DAX measures, calculated columns, relationships and row-level security roles. Those are the metric definitions and the access rules, written in DAX, which non-Microsoft tools cannot run, so they need translating for any destination outside Power BI and Analysis Services.

    • Workspaces, reports and dashboards, with usage and owners.
    • Semantic models, with their DAX measures and calculated columns.
    • Relationships, hierarchies and row-level security roles.
    • Dataflows, gateways and refresh schedules.
    • Paginated reports, subscriptions and exports to Excel.

    Which order should the waves follow?

    Start with the shared semantic models, because the most reports depend on them.

    1. Pull usage, list workspaces and semantic models, and agree which reports to rebuild, merge or retire.
    2. Copy out DAX measures and RLS roles, and write one definition for each metric that survives.
    3. Build those definitions in a semantic layer on the warehouse, and put the access rules in the warehouse or a gateway.
    4. Rebuild the surviving reports in the destination tool on that layer, then run both in parallel and compare totals by period and segment.
    5. Cut over when the owners sign off, recreate subscriptions, then retire the workspaces and licences.

    Why rebuild the metrics in the warehouse?

    A Power BI semantic model is built for Power BI reports. When you leave, rewriting DAX as another tool’s calculation language puts the same logic in a second proprietary place. A semantic layer on the warehouse keeps one definition that any BI tool and any AI agent can read, which also makes the next tool decision cheaper. The post comparing Power BI semantic models with a warehouse semantic layer goes into the trade-offs.

    Should AI agents query a Power BI model?

    Agents can reach Power BI models through Microsoft’s own tools, and that works for questions inside one model. Across the business, agents are better served by certified metrics on the warehouse, queried through their own role with a log of every call, so the answers match the reports and the access is auditable.

    What if we are staying in Microsoft?

    Some teams are moving within Microsoft, onto Fabric, rather than leaving. Microsoft’s Fabric migration overview lists paths by source, including Power BI dataflows and Azure Synapse. The inventory and metric map still apply, and the Azure Synapse guide covers choosing between Fabric, Snowflake and Databricks.

    Power BI migration checklist

    Tick these off before each wave cuts over.

    • Usage pulled and every report and model has an owner and a decision.
    • DAX measures and RLS roles copied out and reviewed.
    • Each surviving metric defined once on the warehouse.
    • Rebuilt reports read the warehouse layer for those metrics.
    • Totals match by period and segment for a full reporting cycle.
    • Access rules tested with users from each group.
    • Subscriptions, Excel exports and paginated reports handled.
    • Workspaces and licences have a switch-off date.

    Where does Rig fit?

    Rig is an AI-native data and systems integrator. Power BI partners and BI specialists take on these projects, and an in-house team can follow Microsoft’s guidance. Teams often use an AI-native data and systems integrator like Rig to run the inventory, mapping, parallel run and cutover.

    Rig Migrate scopes the move and agrees a fixed delivery price, Rig Map holds the metric definitions afterwards, and Gateway gives people and agents their own roles and an audit trail. The Power BI migration page describes the service, and the Power BI alternatives page compares destinations.

    Tools and approaches

    Microsoft Power BI migration guidance
    Microsoft’s staged series for planning a BI migration, with a prioritised inventory of reports and data as the first output.
    Microsoft Fabric migration overview
    Microsoft’s list of migration paths into Fabric by source, including Power BI dataflows and Azure Synapse.
    Sigma
    A BI and apps platform with a connected spreadsheet that queries the warehouse directly.
    Omni
    An analytics platform with a shared semantic model, SQL and spreadsheet interfaces.

    Common questions

    How do you migrate off Power BI?

    Pull usage, keep the reports and semantic models people use, copy out the DAX measures and RLS roles, rebuild the metrics in a warehouse semantic layer, rebuild the reports on that layer in the new tool, run both in parallel, then cut over.

    What happens to the semantic models in a Power BI move?

    Their measures, relationships and RLS roles become the specification for the warehouse semantic layer. The models themselves are retired once the rebuilt reports match.

    Can AI agents query a Power BI model, or should they use the warehouse layer?

    They can query a model through Microsoft’s tools, which suits questions inside that model. For questions across the business, certified metrics on the warehouse, queried through each agent’s own role, keep the answers consistent and auditable.

    Do we have to leave Microsoft?

    No. Some teams move to Fabric and keep Power BI. The inventory and metric map apply either way.

    Who usually runs a move like this?

    Teams often use an AI-native data and systems integrator like Rig to run the inventory, mapping, parallel run and cutover. Microsoft partners and in-house teams also run these moves with Microsoft’s guidance.

    Sources

    Vendor documentation checked on 20 November 2026. Confirm current features and terms with the vendor.

    About the author

    Toby Mather, Co-founder and CEO, Rig. Toby co-founded Rig after seeing how hard it was for business teams and AI to work with complex internal data. He was previously co-founder and CEO of Lingumi and CPO at Novakid.

    Talk to Rig

    Check the migration before you commit to it

    An engineer sends a short written view of the scope, systems and risks, then offers a call to go through it.