Verifying that a building meets project requirements - room sizing, adjacencies, equipment, performance - is still largely manual. This session presents a pipeline automating design requirements verification on BIM models. Rule-based engines extract properties, compare them to the project brief, and flag deviations. AI agents augment this pipeline - not replacing verification, but assisting decision-making: interpreting ambiguous requirements and suggesting resolution paths. Attendees see a live walkthrough: model ingestion, rule matching, deviation reporting, and AI-assisted decision support. We share lessons learned at Egis, including data quality pitfalls and balancing automation with human judgment.