The workflow before#
Bridge Investment Group manages $49.8B in gross assets across more than 30 funds, spanning multifamily, seniors housing, debt and other strategies. Its data lived in separate property management, accounting and investor CRM systems joined by point-to-point feeds. Before each investor report, analysts pulled extracts by hand, reconciled them in spreadsheets and rebuilt numbers they had built the quarter before. There was no master record for a property, fund or investor, so definitions drifted between teams and the same asset could carry different names in different systems. Staff spent close to a third of their time gathering data instead of analyzing it.
What we built#
We built one governed data layer for Bridge, with a single golden record for every property, fund and investor and dashboards that read from it. Investor reports now start from numbers that already agree across systems. We sequenced the work by business value, so the investor reporting team, which felt the pain most, got relief first.
An audit that set the order
We started with a two-month audit of every source system, data flow and reporting need, including interviews with more than 40 business and technical leaders across Bridge's investment divisions. Then we ranked use cases by value: investor reporting first, followed by portfolio performance, property-level reporting and risk monitoring. Ranking up front meant each phase delivered something people used, instead of a long build with a reveal at the end.
Golden records for the core entities
Profisee master data management holds the golden record for properties, funds, investors, vendors and the chart of accounts, with approval workflows for changes. A new acquisition maps into the same structures instead of becoming another silo. We built master data alongside the pipeline rather than after it, because Bridge grows by acquisition and each new deal would otherwise reopen the problem.
One pipeline, five sets of dashboards
Source systems land in Azure Data Lake Storage, Databricks cleans and models the data, and Power BI serves executive, investment, operations, investor and risk dashboards from one governed copy. Reconciliation that used to happen in spreadsheets now runs in the pipeline. Investor report prep takes about 55% less time, and data quality issues in reports are down about 60%.
How we delivered it
Delivery ran in phases: foundation and master data first, then the core pipeline, then the prioritized business use cases. Each phase went live for the teams it served before the next began, with training for the analysts who had been reconciling by hand. Azure Active Directory roles and row-level security controlled who saw what from the first release.
Before and after#
| Measure | Before | After |
|---|---|---|
| Investor report preparation | Several days of manual pulls and rework | 55% less time, from one governed source |
| Cross-system reconciliation | Manual, in spreadsheets | Automated in the pipeline, about 60% less effort |
| Data quality issues in reports | Frequent | About 60% fewer |
| Master records | None; each system kept its own | One golden record per property, fund and investor |
What's next#
With one governed data layer in place, the next step is an assistant that answers portfolio and investor questions directly from the Databricks models, starting with the investor relations team. It will draw on the same golden records, so its answers match the reports investors already receive.




