Seeing Is Believing: AI Observability & Trustworthy AI with SAP & Dynatrace | Webinar Replay

It’s one thing for an AI system to work. It’s another to understand how it’s behaving, why its performance is changing, and what impact its outputs are having downstream. As AI moves into production and becomes embedded in business-critical workflows, observability becomes foundational to trust. Traditional accuracy metrics alone aren’t enough. Teams need visibility into model behavior, input and distributional drift, system performance, and downstream effects to identify issues before they become business problems. In this session, leaders from RaceTrac, SAP, and Dynatrace discuss what meaningful AI observability looks like across the lifecycle—and how it can support governance, risk management, and production reliability. Moderator: • Sean Sanders, Director, Head of Enterprise Data Science @ RaceTrac Speakers: • Shubham Trivedi, Digital Solution Advisor @ SAP • Jason Ostroski, Field CTO @ Dynatrace Sponsors: • SAP • Dynatrace What you'll learn: 1️⃣ Beyond Accuracy Metrics — Understand what teams should monitor when accuracy alone doesn't provide a complete picture of AI health and performance. 2️⃣ Model Behavior Visibility — Explore approaches for identifying drift, distributional shifts, unexpected behavior, and other changes as AI systems operate in production. 3️⃣ Downstream Risk Tracing — Learn how teams can connect AI outputs to downstream applications and business processes to better understand operational impact. 4️⃣ Observability-Driven Governance — See how telemetry and monitoring can support audits, incident response, accountability, and ongoing AI governance. For data science leaders, AI engineers, platform teams, and enterprise architects, this conversation provides a practical look at building the visibility required to operate AI systems with greater confidence. Subscribe for more conversations with enterprise AI leaders and practitioners building reliable, observable, and production-ready AI systems 🧠 Explore upcoming webinars and events: https://datasciconnect.com/events/webinars/