Agents that work out which AI regulations and standards apply to a system, then test whether its controls work, are coming to Collibra's governance platform, following the acquisition of trail ML, a Munich-based company founded in 2023.
trail ML's agents analyse the context surrounding an AI system and determine which frameworks and controls apply. They then assess whether those controls exist and are effective.
The company says this helps organisations operationalise requirements under the EU AI Act, ISO/IEC 42001 and the NIST AI Risk Management Framework.
When supporting evidence changes, assessments can be triggered again. The aim is to replace manual, point-in-time reviews with a continuously updated view of governance and compliance posture.
trail ML also brings runtime controls. These enforce Collibra policies where AI agents run and block actions that would violate them before they happen.
"AI incidents increasingly start with agents taking actions, not models giving answers," Collibra said in announcing the deal.
Context plus automation
Collibra positions itself as an "enterprise AI control plane". It provides the context and control layer for governing data, models, applications and agents.
trail ML adds the automation for analysing evidence, identifying gaps and running governance workflows.
"We founded trail ML because we believe AI governance has to be automated and operational," said Anna Spitznagel, Co-founder and CEO of trail ML.
"For Governance, Risk, and Compliance (GRC) teams, that means less manual work assessing controls and maintaining evidence, and a better way to keep pace with rapidly changing AI environments," said Felix Van de Maele, Co-founder and CEO of Collibra.