AWS Bedrock AgentCore Workshop

​Hands-On Workshop - AI Governance on AWS

​AI agents are transforming how organizations build and operate — but with that power comes a new class of risk. An unprotected agent can be manipulated through prompt injection, leak sensitive data, or take actions no one intended. Governing AI isn't optional anymore; it's foundational.

​Join us for a hands-on workshop that moves AI governance from theory to practice on AWS. You'll see firsthand how an agent can be exploited — and then build the guardrails, controls, and visibility to stop it. By the end, you'll have a working toolkit for securing, monitoring, and continuously verifying AI workloads in your own environment.

​In this workshop, you will:

​Expose the risk — See how unprotected AI agents can be exploited through prompt injection
Build guardrails — Create and apply Amazon Bedrock Guardrails with prompt attack detection, denied topics, and PII filtering
Enforce controls — Use AWS Config rules and conformance packs to continuously verify governance controls
Find the gaps — Inventory AI resources with AWS Resource Explorer and identify governance gaps
Gain visibility — Inspect agent execution traces in Amazon CloudWatch to see agent reasoning and guardrail interventions