Enterprise Data Architecture & Governance for a Large-Cap Telecom
The Problem
A major telecom company's operational data infrastructure had calcified over years of incremental fixes. The core data marts ran on a deeply nested star schema in Oracle, with views stacked on views stacked on views. Any change to the underlying data took one to two weeks to propagate. There was no semantic layer. No data lineage documentation. No formal change management process.
The BI layer consisted of literally thousands of Tableau workbooks, many built by different teams with conflicting metric definitions. A shadow data pipeline had been constructed by a team that couldn't get changes through the official process fast enough. And the institutional knowledge required to navigate all of it was concentrated in a single person who was retiring within weeks.
New executive leadership was demanding daily operational visibility. A separate $30M transformation initiative was underway from a major consultancy, but it was focused on business transformation, not this data architecture problem. The company needed an independent assessment and a buildable plan.
The Solution
We scoped and staffed a 26-week engagement with a three-person team: a fractional Chief Data Officer for strategic architecture and governance, a senior data architect for hands-on technical work, and an AI practice lead for engagement oversight and future-state AI layer scoping.
Phase 1 (weeks 1–4): Current-state assessment. Stakeholder interviews across the data engineering, BI, and operations teams. Full mapping of the existing architecture: Oracle, Snowflake, Informatica, SSIS, a custom Python/Git DAG tool, and the shadow Data 360 pipeline. Documentation of the institutional knowledge at risk of walking out the door.
Phase 2 (weeks 5–16): Reference architecture design. A modern medallion architecture layered on top of the existing Oracle/Snowflake coexistence (both platforms were staying, and the engineering team had deep Oracle expertise they weren't going to abandon). Tool recommendations for automated data lineage, documentation, and DDL generation. A consolidation plan for the shadow pipeline. Semantic layer design for business-user self-service, reducing dependency on the massive Tableau workbook sprawl.
Phase 3 (weeks 17–26): Implementation roadmap and handoff. Executable plan calibrated to the pace the new executive leadership expected. Evaluation framework for the offshore contractor resources already standing by to begin execution. Governance framework with change management processes to prevent the architecture from re-calcifying.
The Result
The engagement delivered a buildable reference architecture that the client could present to leadership and use to direct development resources. The knowledge transfer was completed before the retiring team member departed. The governance framework introduced formal change management for the first time, and the semantic layer design created a path toward self-service analytics that would reduce the BI team's workbook maintenance burden. The architecture was designed to support an AI and automation layer as a future expansion once the data foundation matured.