The challenge
A global manufacturer of construction tools wanted its sales teams to manage business opportunities proactively, powered by AI-driven recommendations. That meant building an analytics foundation reliable enough for AI chatbots to answer real business questions accurately, and flexible enough to keep pace with new use cases as they emerged.
What we did
Designed a Bronze, Silver & Gold data architecture in Microsoft Fabric, aligned with data federation principles and meeting requirements for quality, security, governance and confidentiality.
Built an AI-ready semantic layer, modelling business entities, metrics and relationships so AI-powered chatbots could answer analytics questions with consistent, context-aware definitions.
Introduced an AI-assisted engineering workflow, using Claude and VS Code to generate design specs and mentoring the team on AI-collaborative development practices.
Developed PySpark and Spark SQL pipelines to ingest and transform source data spanning customer meetings, opportunities, app telemetry and recommendation adoption tracking.
Established automated data quality controls and test frameworks, catching completeness, validity and consistency issues before data products were published.