Microsoft Research has described an Industry Evaluation Data Generation Platform designed to test Microsoft Copilot against realistic enterprise scenarios without using customer data. The company says the platform creates simulated organizational environments containing users, meetings, emails, documents, reporting structures and evolving project histories.

According to Microsoft, the synthetic environments are grounded in public sources including the U.S. Department of Labor’s O*NET occupational database, online job postings and SEC filings. These sources are used to model employee roles, reporting structures, departmental compositions, organizational scale and industry workflows. Microsoft says its legal and policy organization reviewed and approved the sources.

The platform currently includes synthetic environments representing pharmaceutical, capital markets and IT/software organizations. Microsoft says the environments include activities such as clinical-trial discussions, compliance communications, portfolio reviews, sprint planning and product-launch coordination.

An automated six-stage pipeline generates and validates industry-specific Copilot queries from generic workplace tasks. The company evaluates the resulting data for structural fidelity, contextual relevance, behavioral realism and content safety. Microsoft says the datasets contain no real customer data and that automated safety checks screen generated content for harmful or personally identifiable material.

Microsoft also says automation reduced dataset creation from a process taking multiple days to automated runs. The company estimates savings of about $100,000 per evaluation dataset and $300,000 across the project. These figures are Microsoft’s estimates and were not independently verified in the source.