Democratizing Data and AI at Freshworks With Governed Lakehouse and Freddy AI

Freshworks has standardized its data and AI strategy on a Databricks‑powered lakehouse called Baikal, built to serve 70,000+ customers and 1,500 internal users across 14 business functions. Databricks ingests data from dozens of operational and third‑party sources into Delta Lake bronze, silver, and gold layers governed by Unity Catalog, enabling high‑quality, discoverable datasets and fine‑grained access control with full auditability. Dedicated BI, AIML, product and functional workspaces run on governed & segregated compute with centralized monitoring, improving reliability, observability and cost attribution. On this foundation, Freshworks trains and operates Freddy AI agents, copilots and insights using Databricks ML, MLflow and an open AI ecosystem to move seamlessly from experimentation to production. This architecture modernizes legacy Hadoop systems, delivers 2–5x faster processing and democratizes AI across the company.