Intent
AI reads the estate: packages, procedures, reports, pipelines. What each one is for and what depends on it comes out as a graph, not a guess.
Service 04 / 05 · Databricks and Microsoft
EstateMigration.
One method, three destinations. Databricks: Synapse, SSIS and T-SQL estates become the Lakehouse. Microsoft: SSRS reports become Power BI, and SQL Server, SSIS and Synapse estates become OneLake on Fabric. In every case AI recovers intent, deterministic generators produce editable artifacts, and an engineer approves every consequential decision.
01 / Choose the destination
Three tracks
Microsoft · two lanes
0203
To Power BI
SSRS estates become Power BI projects: the logic inside every RDL recovered, the right target chosen per report, editable PBIP, TMDL and paginated artifacts generated, every report validated against its original.
To Fabric
SQL Server, SSIS, ADF and Synapse estates become OneLake: mirrored where a source stays, rebuilt as Fabric pipelines, Warehouse and Lakehouse items where it moves, Direct Lake models under the same reports.
02 / One method
Five gates, every object
Target
Each object gets a destination: rebuilt, consolidated, served by a model, or retired. An engineer decides; the reason is recorded.
Build
Deterministic generators produce the artifacts, editable and in Git. Nothing opaque, nothing you cannot change.
Evidence
Every object is validated against its original: figures, rows, behaviour. The evidence travels with it.
Cutover
Consumers repointed, the original retired, the runbook handed over. Approved by a person, every time.