Airia launches Model Risk Management solution integrated with Microsoft Foundry to enable continuous validation of autonomous AI agents

Airia announced availability of its Model Risk Management (MRM) solution integrated with Microsoft Foundry on May 26, 2026, delivering continuous validation, automated compliance reporting, and governed improvement workflows for organizations deploying autonomous and auto-improving AI agents (systems that refine their own behavior through feedback loops and self-directed optimization).

The solution orchestrates Foundry's evaluators for agent quality, safety, and NLP metrics within Airia's governance layer, with scheduled and event-driven evaluation runs against production traffic; it also detects behavioral drift when agents deviate from expected performance parameters and integrates with Microsoft Copilot Studio agents, Foundry models, Microsoft Defender for AI, and Microsoft Purview.

The MRM solution addresses the challenge that enterprises report pervasive shadow AI—unsanctioned AI tools deployed by employees that create data privacy risks, compliance violations, and operational blind spots—while some regulations now mandate that organizations validate, monitor, and document everything flowing through an AI model.