Red Hat has launched an open-source community project called asago (AI Safety and Governance Orchestration), bringing together IBM, Microsoft, Brave Software, MIT and The Alan Turing Institute to turn written AI governance policies into automated, deployment-ready safety controls across hybrid cloud environments.
Bridging Policy and Deployment
Rather than relying on generic technical tooling, asago is designed to work directly from an organisation's own custom policy documents, translating governance language into measurable risk profiles and deployable safety controls. The platform can read uploaded AI governance policies and convert that language into actionable risk assessments.
Red Hat points to frameworks such as the EU AI Act, via the IBM AI Risk Atlas, and the NIST AI Risk Management Framework as examples of the kinds of policy documents the system may be able to interpret. Rather than testing against generic benchmarks, asago will generate scenarios tailored to specific use cases, then recommend mitigations and orchestrate them into deployment-ready configurations complete with audit trails.
"asago intends to holistically smooth and streamline how AI safety controls are operationalised within an enterprise organisation, not with the technical tooling, but with the organisation's own custom policy documents."— Stuart Battersby, AI Safety & Model Evaluation Architect, Red Hat
From AI Pilots to Autonomous Agents
According to Steven Huels, Vice President of AI Engineering at Red Hat, operational guardrails have become critical infrastructure as organisations shift from experimental AI pilots toward autonomous agents running in production. He frames asago as a complement to Red Hat's existing Lightwell initiative, which focuses on securing the open-source supply chain from AI-driven vulnerabilities.
The goal is to give compliance teams, data scientists and infrastructure administrators a single open standard to work from, reducing the inconsistencies that manual translation introduces when converting abstract policy guidelines into operational, engineering-ready controls that fit within DevOps and GitOps workflows.
Open for Community Participation
asago is currently in its project formation phase on GitHub, with developers, academic researchers and enterprise early adopters invited to view the repository and take part in project governance. Red Hat is specifically encouraging collaborators from a wide range of global jurisdictions to ensure the project captures diverse AI safety viewpoints and regulatory perspectives.