22-25 April 2026

How is Microsoft Foundry like Pokémon?

Proposed session for SQLBits 2026

TL; DR

Gotta Catch 'Em All: Building Your AI Agent Team with Microsoft Foundry Every Pokémon trainer knows you don't win with just one overpowered Charizard—you need a balanced team of specialized agents working together. Microsoft Foundry is your Pokédex and Battle System combined, helping you select, train, and orchestrate AI agents that complement each other's strengths. In this session, I'll share real-world battle scars from building agentic AI solutions with Microsoft Foundry across charity fundraising, medical research, and supply chain optimization.

Session Details

Gotta Catch 'Em All: Building Your AI Agent Team with Microsoft Foundry
Every Pokémon trainer knows you don't win with just one overpowered Charizard—you need a balanced team of specialized agents working together. Microsoft Foundry is your Pokédex and Battle System combined, helping you select, train, and orchestrate AI agents that complement each other's strengths.

The AI and Agentic AI stack is evolving quickly, with options from different vendors. In the blue corner, we now have Microsoft Foundry (formerly AI Foundry) as new kid on the block

In this session, I'll summarise real-world battle scars from building agentic AI solutions with Microsoft Foundry across charity fundraising (donor engagement and impact networks), medical research (literature analysis), and supply chain optimization (logistics and constraint management).

3 things you'll get out of this session

- Understand what Microsoft Foundry is and when to use it for agentic AI solutions - Learn how to architect multi-agent systems with specialized AI agents working together - Master patterns for agent orchestration, tool use, and autonomous task completion - Integrate Foundry with Fabric, LLMs, and RAG for production deployments - Identify and avoid common pitfalls: hallucinations, infinite loops, and cost overruns - Apply decision criteria to determine when Foundry is appropriate vs alternative approaches