What happens when AI agents stop working alone and start collaborating?

WHAT YOU'LL BUILD

Explore agentic AI through a real-world use case: multi-agent stock analysis.

Starting from a working swarm of specialized agents, you'll extend the system by adding new capabilities, introducing new agents, and shaping how they coordinate.

You'll gain hands-on experience:

Creating tools for agent collaboration
Designing handoff chains between agents
Orchestrating multi-agent systems
Building specialized agents that work together to solve complex problems

Walk away with practical experience building collaborative AI systems that go beyond single-agent limitations.

KEY TOPICS

Multi-agent system architecture
Agent specialization and collaboration patterns
Tool creation for agent ecosystems
Handoff chain design
Swarm orchestration strategies

WHO SHOULD ATTEND

AI/ML engineers, developers building agent systems, technical leads exploring multi-agent architectures, and builders interested in collaborative AI.

PREREQUISITES

Participants must complete setup before the workshop:

Required Software:

AWS account with Amazon Bedrock model access enabled (Claude Opus 4.6 via cross-region inference)
AWS CLI v2 installed and configured (aws configure)
Python 3.12+ and uv package manager installed
Claude Code CLI installed

Setup Steps:

Add the Strands Agents MCP server to Claude Code: claude mcp add strands uvx strands-agents-mcp-server
Clone the workshop repo:

git clone --depth 1 --filter=blob:none --sparse https://github.com/strands-agents/samples.git
cd samples
git sparse-checkout set python/04-industry-use-cases/finance/finance-assistant-swarm-agent
cd python/04-industry-use-cases/finance/finance-assistant-swarm-agent

Install dependencies: uv sync
Verify setup: uv run finance_assistant_swarm.py
- Enter a ticker (e.g., AMZN)
- Wait for the report to complete
- If a full analysis report appears without errors, you're all set

WORKSHOP CONDUCTORS

WORKSHOP ACCESS

This workshop is exclusively available to summit attendees. First come, first served with limited capacity.

[Hands-on Workshop]From One Agent to a Swarm: Building Collaborative AI Analysts