ESPC26
Conference Sessions
Copilot Studio, Microsoft Foundry, or Both? Choosing the Right Architecture for Intelligent Agents
SPEAKERS
Katerina Chernevskaya
MVP
AI Cloud Agency
Maria Dicheva
KPMG Delivery Network
Bülent Altinsoy
MVP
Avanade
ABOUT THE SESSION
Building intelligent agents is no longer the hard part.
The real challenge is deciding which technology to use, for which problem, and why.
This hands-on workshop focuses on a single, practical question:
When should you use Copilot Studio, when should you use Microsoft Foundry, and when does it make sense to combine both?
Participants work through the same business scenario using different approaches:
implementing the solution with Copilot Studio as part of a business workflow (Work IQ)
implementing the same logic with Microsoft Foundry as a dedicated intelligence layer (Foundry IQ)
combining both into a single, well-structured architecture
By building and comparing these approaches side by side, participants gain a clear, experience-based understanding of:
- the boundaries of low-code agents
- the strengths and responsibilities of a Foundry-based AI layer
- the trade-offs between speed, control, governance, and maintainability
The workshop is practical, architectural, and decision-oriented.
No hype, no generic AI promises, and no “one tool fits all” answers.
Who should attend:
This workshop is designed for architects, technical leads, and Power Platform professionals who need to make informed, defensible technology choices when designing intelligent solutions.
Participants leave with concrete criteria they can apply immediately in real projects, not just another demo.
Assumed Knowledge:
Solid understanding of Power Platform, Copilot Studio, and basic Azure AI concepts; experience with solution architecture and ALM practices.
Practical Takeaways:
Clear decision framework: Copilot Studio vs Foundry vs hybrid. Hands-on experience building the same solution in different architectures. Understanding trade-offs (speed, control, governance, maintainability). Practical patterns for combining low-code and AI layers. Criteria for making defensible architecture decisions.
Out of Session Scope:
Beginner-level introductions, deep custom model training, non-Microsoft AI stacks, and theoretical architecture discussions without implementation.
MEET THE SPEAKERS
Katerina Chernevskaya
MVP
AI Cloud Agency
Maria Dicheva
KPMG Delivery Network
Bülent Altinsoy
MVP
Avanade