Building the data and AI backbone of a global broadcast group
A multi-year Centre of Excellence on Microsoft Azure and Databricks — three pillars, one canonical foundation, and AI running in production.
Where the engagement started
A global broadcast and media services organisation, operating across regions and grown through acquisition, needed one dependable picture of its business — and a foundation ready for AI. Reporting was fragmented across ERP, CRM and finance systems; the same business questions produced different answers in different regions. The question we started with: what must this platform make possible for the group's leadership, operations and AI ambitions — in that order?
What we designed and delivered
A three-pillar Centre of Excellence
One operating model spanning BI, operations and AI — so the platform serves the whole business rather than a single reporting need, with governance and resourcing designed in.
A canonical enterprise model
Conformed dimensions and composite business keys unifying ERP, CRM and finance sources across regions — every report and model drawing from one agreed truth.
Governed lakehouse engineering
Medallion architecture on Databricks with Unity Catalog — full lineage, schema-drift-resilient ingestion, and CI/CD discipline throughout.
Agentic AI in production
LLM document-intelligence pipelines running in production, with structured extraction, confidence scoring and human-in-the-loop review — the AI pillar delivering, not demonstrating.
How the value shows up
One version of the truth
Group-level reporting on conformed data — the reconciliation debates replaced by a shared foundation.
AI with a bloodstream
Production AI fed by governed, lineage-traced data — dependable enough to act on.
Capability that stays
Architecture decisions documented and transferred — the client's team grows with the platform.
Back to selected work.