Addepar, a financial data platform managing nearly $9 trillion in assets, built an AI assistant called Addison on Databricks running on AWS. The system uses Databricks Agent Bricks, Foundation Model APIs, Unity Catalog for governance, and MLflow for lifecycle management to deliver agentic workflows for investment professionals. Key outcomes include 60% pipeline cost reduction, $2M+ in infrastructure savings, and 5x faster pipeline delivery. Addison moves beyond simple LLM prompts to multi-step, governed agentic workflows including portfolio analysis, automated insight generation, and Supervisor Agent coordination with Genie-powered analytics.
Table of contents
Safe, scalable GenAI for financial servicesFrom LLMs to agents with Agent Bricks, Foundation Model Serving and MLflowCollaboration and data sharing as a force multiplierA partnership focused on outcomesSort: