ENTERPRISE DATA & FINOPS
CLIENT
Media & Intellectual Property Rights Entity

Illustrative UI mockup. Real client data and proprietary schemas are sanitized for confidentiality.
Modern Fabric Lakehouse architecture with automated FinOps compute cost control
Data & AI delivery engineered for enterprise clarity.
PROJECT STORY
The Business Challenge
The client, a media and intellectual property rights entity, faced the complex task of unbundling and centralizing multiple disconnected systems—including an on-premise PHC ERP, an Outsystems enterprise portal, and external data feeds (Dashmote and GfK)—into a unified source of truth.
The organization faced key architectural and operational hurdles:
Fragmented Revenue Operations: Disconnected visibility between operational activity (contracts, on-site inspections, licensing) and core financial events (invoicing, collections, legal actions).
Uncontrolled Cloud Overhead: The necessity to build modern lakehouse pipelines while strictly eliminating idle cloud compute costs.
Complex Entity Resolution: High volumes of inconsistent entity naming across disparate operational systems requiring predictive resolution and automated churn forecasting.
Fragile Ingestion Pipelines: Ingestion routines vulnerable to external API schema drifts and intermittent connection failures.
The Technical Solution
Sersight architected a robust Medallion Architecture (Bronze, Silver, Gold) on Microsoft Fabric, orchestrating resilient ingestion, machine learning, and dynamic FinOps automation:
FinOps Cloud Automation (Azure Automation & Fabric F-SKUs): Engineered an intelligent orchestration pipeline that scales compute capacity on-demand—triggering F-SKUs only during active extraction and semantic model refreshes, and shutting down compute immediately upon job completion to minimize total operational costs.
Metadata-Driven Ingestion Engine (PySpark & OneLake): Built a self-healing, configuration-driven ingestion engine capable of automatically adapting to upstream schema drifts and recovering from transient network failures.
Embedded Machine Learning & Entity Reconciliation: Integrated automated Python ML routines into Delta Lake pipelines, employing morphological fuzzy matching to resolve cross-system records and injecting predictive churn scores directly into fact tables.
High-Density SVG Visual Engineering (Power BI / DAX): Replaced default visuals with custom-coded SVG elements in DAX, rendering pixel-perfect financial execution rates, YTD vs. LY comparisons, and net portfolio trends into an executive interface.
The Strategic Impact
The unified platform transformed the client’s data landscape into an automated, cost-efficient analytics ecosystem:
Minimal Cloud Compute Spend: Zero idle infrastructure costs through automated, event-driven compute scaling.
End-to-End 360° Revenue Traceability: Complete transparency across the entire revenue lifecycle, tracking every licensing euro from field audit to cash collection.
Predictive Decision Intelligence: Transitioned leadership from reactive descriptive reporting to proactive risk management and portfolio forecasting.
MORE WORK
