Banks building AI agents on Databricks can now point them at ActiveViam's risk engine rather than have a language model improvise SQL against raw trade tables. ActiveViam, which sells risk analytics to capital markets firms, has announced a collaboration with Databricks that exposes its calculations as tools an agent can call.
The tools sit in Atoti Intelligence, the agentic architecture of ActiveViam's Atoti platform, and cover value at risk, expected shortfall, FRTB, xVA and the LCR and NSFR liquidity ratios. Each calculation is deterministic and runs in the customer's own environment under its existing Databricks governance, lineage and access controls. The institution supplies its own approved language model, so data and intellectual property stay on the platform, ActiveViam said.
What changes for each customer base
The integration builds on ActiveViam's DirectQuery support for Databricks, which already queries governed Databricks data live. Existing ActiveViam customers can move onto Databricks without rebuilding risk: measures, hierarchies and workflows carry over, and the same definitions serve Databricks-native agents and tools. Databricks customers can add capital-markets risk analytics to their lakehouse without building it in-house or moving data off the platform.
The pitch is repeatability. A risk number that an agent assembles from generated code can differ from run to run; one produced by a fixed calculation can be traced to the trade and signed off. Alan Lee, ActiveViam's Chief Product Officer, put it that way.
“Databricks gives banks one governed home for their data, while Atoti turns it into risk intelligence they can trust and sign off on, without copying data out,” said Alan Lee, Chief Product Officer at ActiveViam.
ActiveViam gave no customer names, pricing or availability date.
To stay across the latest in cloud, AI and enterprise tech analysis from Compare the Cloud, subscribe to our weekly newsletter at https://www.comparethecloud.net/newsletter
