Fortune 500 tech company unified HR data and gave managers a GenAI layer to ask questions in plain language.

A Fortune 500 consumer technology company whose HR organization supports a workforce spanning six continents, from retail staff to specialized engineering teams. Leadership wanted people analytics at the same rigor they applied to product and supply chain decisions, but the data was fragmented and metrics were defined inconsistently across teams.

People data lived in multiple source systems: HRIS, payroll, recruiting tools, performance management, learning. Each had its own definitions and quirks, and manager in one system didn't always match manager in another. Time-to-fill was calculated three different ways depending on which team you asked. Active headcount on a given Friday could pull a different number from each system. Getting a clear answer to a basic leadership question could take days, and confidence in the answer was capped by the team's confidence that the underlying definitions were sound.

They had the infrastructure for one of the largest private Tableau deployments anywhere, but having the infrastructure wasn't the same as having a system people trusted. Every data pull carried strict privacy requirements that varied by jurisdiction (GDPR, country-specific labor law, internal policy), so quick-and-dirty solutions weren't an option. And the manager population was big enough that any inconsistency multiplied quickly: a definition that drifted between two reports turned into a thousand half-correct status meetings the next month.

The Blueprint ran six weeks: stakeholder interviews across HR, recruiting, performance, and the regional people teams; a full audit of every source system and every metric definition currently in production; a privacy review across the jurisdictions in scope. We came out of Blueprint with about a hundred conflicting metric definitions to resolve and a sequenced roadmap for which dashboards mattered most across the operating cadence.

Through the four-month MVP and the multi-year expansion that followed, we ran weekly working sessions with the people analytics team and monthly steering with HR leadership. Dashboards shipped in cohorts every four weeks rather than waiting for a single big launch. The dbt project ran as a shared codebase from week three onward: their team and ours committing into the same repo, with code review running through their normal engineering process.

Knowledge transfer was the engagement model, not a phase. The people analytics team was co-owner from week one. By the end of the MVP, they were merging the majority of model changes themselves. The expansion phases scaled the team rather than replacing it. When the GenAI layer launched, the prompt library was owned by their analytics team and updated alongside the dbt project, so the conversational layer never drifted from the underlying source of truth.

• Consistent metric definitions across every HR source system, documented and tested in dbt
• 50,000+ monthly dashboard views across HR, recruiting and people-manager workflows
• 40+ Tableau dashboards in active use across specialized workflows
• GenAI layer putting answers directly in managers' hands without waiting for an analyst
• Privacy controls built into the data models rather than bolted on top
• Initial scope grew into a multi-year strategic partnership across business units