Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
This checklist provides a structured framework to initiate productive architectural discussions and ensure all critical aspects are considered before you begin the design of an agentic solution.
Use this checklist to:
- Establish foundational requirements across governance, user experience, and technical specifications.
- Identify potential challenges early in the planning process to avoid costly redesigns.
- Align stakeholders on key architectural decisions and constraints.
- Ensure comprehensive coverage of security, compliance, and cost considerations.
Each section includes actionable guidance and links to relevant resources to help you make informed architectural decisions. Work through these items with your team before moving on to detailed solution design.
| Category | Details | Learn more |
|---|---|---|
| Acquisition and governance | Understand governance, change management, cost, and acquisition metrics to establish the framework for sustainable agent deployment. This understanding includes evaluating organizational readiness for agent adoption and integration with existing governance processes. | Microsoft 365 Copilot Agent Management Essentials |
| Entry points and user experience | Determine how interactions start, including client options, user experience patterns, and trigger mechanisms. Make sure agents integrate effectively with existing workflows and user expectations. | Design review checklist for Experience Optimization |
| Processing and data flow | Define logical flows, involved tools, and data retrieval requirements throughout processes. This information enables accurate assessment of technical complexity and resource requirements. Include threat modeling and security analysis to identify potential vulnerabilities. | Threat Modeling - Engineering Fundamentals Playbook Data Security and Compliance for GenAI |
| Input and output specifications | Identify prompt characteristics, data inputs including size, length, data types, and context size requirements. This information ensures agents can handle expected workloads. | Observability in generative AI with Microsoft Foundry |
| Output specifications | Clarify the anticipated content, its format, delivery channel, and user feedback mechanisms. | Observability in generative AI with Microsoft Foundry |
| Control and accountability | Specify mechanisms for audit, traceability, and security to maintain accountability and safeguard solution integrity. This process includes establishing monitoring and evaluation frameworks for ongoing operational assessment. | Security design principles - Azure Well-Architected Framework Security design principles - Power Platform Well-Architected |
| Return on investment opportunities | Assess return on investment opportunities and understand how Microsoft 365 Copilot licensing influences design choices. Ensure architectural decisions align with economic constraints and organizational licensing strategies. | Understanding the RICE Model and its framework – Microsoft 365 Copilot Studio agent usage estimator (preview) Pricing Calculator | Microsoft Azure Copilot Studio licensing Licensing and cost considerations for Copilot extensibility options |
Next step
Learn about the components that work together to enable agents to receive user input, process requests, access external resources, and deliver meaningful responses.