Case study · Canadian bank
Two billion records into a governed data model.
A major Canadian bank needed its IT asset and request data, scattered and unstructured, brought into a modeled environment that applications and automation could safely build on.
The challenge
Asset information existed as massive unstructured datasets, over two billion rows in CSV extracts, with no schema, no relationships and no role-based control. Every application built on top inherited that fragility.
The approach
We composed the data schema first: tables, columns and relationships designed before any data moved. The datasets were loaded into Microsoft Dataverse using dataflows built to handle the volume, and a Power Apps application was developed on the governed model for request entry, summary and submission, with Power Automate handling downstream processing based on the data entered.
The outcome
The bank moved from spreadsheet-scale chaos to a relational, secured data model holding billions of rows, with applications and automation operating on governed data instead of raw files. The schema-first approach meant integrity and access control were properties of the platform, not hopes about user behavior.