Nearly every business is constantly trying to identify, analyze, and grow their market. Yet, traditional processes for data management inevitably lead to a database that contains missing, outdated, invalid, or inconsistent data. Automated knowledge graph construction techniques are a scalable way to maintain and enforce the quality of data in real-world business databases and knowledge graphs offer an ideal representation for modeling the complex relationships between business entities and for validating properties and constraints. In this presentation, we will cover lessons learned building a commercial system that maintains 250M business entities and automatically updates over 100M new facts each month as the external world changes.