Freddy AI at Scale: How Freshworks Democratized Data Intelligence for 70,000 Customers on a Governed Lakehouse

When your company builds AI software for 70,000 customers, the quality of your own internal data infrastructure is not a technical detail — it is a competitive differentiator. Freshworks standardized its data and AI strategy on a Databricks-powered lakehouse. Built on Delta Lake with Unity Catalog governance, data is ingested from dozens of operational and third-party sources into bronze, silver, and gold layers — enabling high-quality, discoverable datasets with fine-grained access control and full auditability. On this foundation, Freshworks trains and operates its Freddy AI agents, copilots, and insights engine, achieving 2–5x faster processing while replacing legacy Hadoop systems. In this session, Sreedhar Gade and Prem Kumar Patturaj reveal the architectural decisions that made it possible to democratize AI across the enterprise while maintaining the governance standards that regulated enterprise customers require. For CTOs and CDOs who need AI infrastructure to support both internal operations and external product commitments, this session delivers the blueprint.