Key takeaways
- The checklist does not change with the tool. Power BI, Tableau, Looker, Qlik, SSRS and Cognos moves all follow inventory, metric map, parallel run and cutover.
- Usage data decides the scope, so rebuild the reports people open and retire the rest once their owners agree.
- The metric map is the step that keeps the old and new numbers the same, and it belongs in a warehouse semantic layer that every tool can read.
- A parallel run compares business totals and permissions as well as how the dashboards look.
- Cutover happens one wave at a time, with a named owner signing off each wave.
- 01Inventory
Reports, owners and usage
- 02Map metrics
Into a warehouse semantic layer
- 03Parallel run
Totals and access compared
- 04Cutover
Wave by wave, with sign-off
What are the four steps?
Every BI migration we run follows the same plan, and the source tool only changes the details of each step. The order matters because each step produces the input for the next one.
- Inventory the reports, dashboards, data sources and scheduled deliveries, with an owner and a usage figure for each.
- Map every metric that survives the inventory to one definition in a semantic layer on the warehouse.
- Rebuild the reports in the destination tool on that layer, then run old and new side by side and compare the numbers.
- Cut over one wave at a time when the owner signs off, then switch off the old reports and licences.
What should the inventory capture?
Most BI tools can tell you who opened what and when, either in an admin log or a usage report. Start there, because it turns a political argument about which reports matter into a list with dates on it.
Capture the logic as well as the objects. Calculated fields, filters baked into a dashboard and row-level rules are where the business definitions hide.
- Reports and dashboards, with views in the last 90 days and a named owner.
- Data sources, extracts and the warehouse tables behind them.
- Calculated fields and measures, with the formula copied out.
- Row-level security rules and the groups they apply to.
- Subscriptions, exports and emailed reports that people rely on without opening the tool.
Why does the metric map go in the warehouse?
If you rebuild each metric inside the new BI tool, you have moved the problem to a new place. The next tool change, or the first AI agent that queries the warehouse, will find the same metric defined three ways.
Put the definition in a semantic layer on the warehouse instead. Snowflake semantic views, Databricks metric views, dbt’s semantic layer and Rig Map all do this in different ways. The destination BI tool then reads the definition from the warehouse layer.
What do you check in the parallel run?
Compare the numbers the business uses to make decisions, at the grain they use them. A revenue total that matches for the year can still be wrong for every month, so compare by period and by segment.
Test permissions with real users from each group. A rebuilt report that shows a sales rep every region has failed the parallel run, however good it looks.
The checklist
Tick each item per wave before you switch anything off.
- Usage data pulled and every report has an owner.
- Reports marked rebuild, merge or retire, and owners have agreed.
- Every surviving metric has one definition in the warehouse layer.
- Rebuilt reports read the warehouse layer for those metrics.
- Totals match by period and segment for a full reporting cycle.
- Row-level rules tested with users from each group.
- Subscriptions and exports recreated or retired.
- Owner sign-off recorded, and the old report has a switch-off date.
What does this look like in practice?
Birdie, which makes care management software, replaced its legacy BI dashboards with Rig. Its metric definitions and dbt models moved onto Birdie’s own Snowflake as certified metrics, and the RevOps lead who had spent about three months building the old reporting said they rebuilt virtually all of the reports in less than a day. The full story is in the Birdie case study.
Which tool guide should I read?
The legacy BI migration guide covers Cognos, SAP BusinessObjects and MicroStrategy. There are separate guides for Power BI, Tableau, QlikView, SSRS and Crystal Reports, and Looker, each with the vendor’s own tooling described from official pages.
Where does Rig fit?
Rig is an AI-native data and systems integrator. A systems integrator can staff the programme, or an in-house team can follow the destination vendor’s official tools. Platforms like Rig handle mapping, parallel runs and cutover, and leave a governed semantic layer on the warehouse you already run.
Rig Migrate runs the inventory and the rebuild at a fixed price, Rig Map holds the metric definitions afterwards, and Gateway controls which people and AI agents can query them. The BI migration page sets out the service.
Tools and approaches
- Microsoft Power BI migration guidance
- Microsoft’s migration series describes five stages for moving from another BI tool to Power BI, starting with a prioritised inventory of reports and data.
- Snowflake semantic views
- A schema-level object that defines metrics, dimensions and relationships, used by Cortex Analyst, Cortex Agents and BI tools.
- Databricks metric views
- Metric views in Unity Catalog define business metrics once and are read by dashboards, Genie and external BI tools such as Power BI, Tableau and Sigma.
Common questions
What is a BI migration checklist?
It is the list of checks for moving reports and metrics off one BI tool: an inventory with owners and usage, one warehouse definition per metric, a parallel run that compares totals and permissions, and a signed-off cutover per wave.
Is this different from the legacy BI migration guide?
Yes. This page is the shared set of steps for any tool. The legacy BI guide is the yearly guide with sections for Cognos, SAP BusinessObjects and MicroStrategy.
What should we inventory before a BI move?
List the reports and dashboards with their usage and owners, the data sources behind them, calculated fields with their formulas, row-level security rules, and the subscriptions and exports people rely on.
When do we cut over from the old reports?
Cut over a wave when its totals have matched for a full reporting cycle, permissions have been tested with real users and the report owner has signed off.
Who usually runs a BI migration?
Platforms like Rig handle mapping, parallel runs and cutover, and leave a governed semantic layer on the warehouse you already run. An in-house team or a systems integrator can run the same checklist.
Sources
Vendor documentation checked on 20 November 2026. Confirm current features and terms with the vendor.