As enterprises expand their use of AI models, applications, agents, and autonomous workflows, organizations are increasingly leveraging AI models to monitor, test, evaluate, and govern other AI systems. These supervisory AI models can automate quality assurance, detect errors and hallucinations, validate outputs, monitor performance, identify risks, and enforce governance policies.

However, building, deploying, and managing AI models responsible for overseeing other AI systems introduces significant architectural, engineering, operational, and business challenges. How can enterprises ensure these supervisory models are accurate, reliable, scalable, and cost-effective while maintaining appropriate human oversight and accountability?

Topics of discussion will include, yet will not be limited to:

• Architecting and integrating supervisory AI models into enterprise AI environments
• Building, training, testing, and managing AI models that oversee other AI systems
• Automating quality assurance, output validation, error detection, and performance testing
• Continuously monitoring AI applications, agents, and workflows for anomalies, model drift, and operational risks
• Establishing AI-driven governance, security, compliance, and human oversight mechanisms

Michael Arulfo, Chief AI Officer, CADRE

Marty Samadder, Head of Global Enterprise Architecture, NOVELIS

Laxmikant Pukale, Director of Advanced Capabilities - AI & Intelligent Automation, USAA

Moderator: SriHarsha Pushkala, Director, Fraud Strategy & Analytics, ATLANTICUS