Agentic AI on AWS: Bedrock + API Gateway 3.12 MCP Proxy

_Video_

Use Amazon Bedrock Agents, Kong’s MCP proxy plugin, and AI guardrails to move from POCs to secure, observable agentic systems at enterprise scale.

An AWS Partner Solutions Architect outlines how to operationalize agentic AI using Agents for Amazon Bedrock, and where Kong Gateway 3.12 fits with MCP proxy, observability, and guardrails.

What you’ll learn:

- Agentic AI fundamentals: tools, goals, memory, observability

- Scaling with Agents for Amazon Bedrock (serverless, MCP gateway, identity, OpenTelemetry)

- Model strategy: Amazon Nova + third‑party models via Bedrock

- Governance: Bedrock Guardrails + Kong AI guardrails (prompts, cost, policy)

- Kong Gateway 3.12 MCP proxy: expose API actions to agents safely

- AI observability: token usage, rate limits, per‑team controls

**Topics**

- [Agentic AI](/resources/videos/tag/agentic-ai)Agentic AI- [AWS](/resources/videos/tag/aws)AWS- [Kong Gateway](/resources/videos/tag/kong-gateway)Kong Gateway- [API Summit](/resources/videos/tag/api-summit)API Summit

# Ship Agentic AI on AWS with Kong Gateway 3.12

## Get started with the unified API and AI platform

## step-0