Databricks agents
Monte Carlo's Agent Trust gives you self-improving Databricks agents. Bayes finds failures in Agent Bricks, Mosaic AI, and Genie conversations and proposes a verified fix.
Monte Carlo's Agent Trust gives you self-improving Databricks agents, including Agent Bricks Knowledge Assistant templates, custom Mosaic AI Agent Framework agents, and AI/BI Genie spaces. Bayes, Monte Carlo's AI teammate, finds what's going wrong in real user conversations, traces each problem to its cause, and hands your team a fix to review. Monte Carlo reads every type through your existing Databricks SQL Warehouse connection. Only how traces reach Monte Carlo differs by agent type.
Choose your agent type
- Agent Bricks and custom agents — Knowledge Assistant templates and custom Mosaic AI Agent Framework (MLflow SDK) agents. They log MLflow traces to a Unity Catalog table once you enable trace storage, and Monte Carlo reads that table.
- AI/BI Genie — Genie spaces. Their conversation history is API-resident, so Monte Carlo installs a managed collector that materializes it into a Unity Catalog table, then reads that table.
Prerequisites (all agent types)An existing Databricks integration in Monte Carlo with a SQL Warehouse connection. If you haven't set this up yet, see our Databricks SQL Warehouse guide.
All types connect through the same Settings → Agent Observability → Add → Platform Agent flow — follow the guide for your agent type above.
Updated 16 days ago
