This module extends topic architecture into the action and flow design layer, prompt action design, multi-turn conversation flow design with variable management, and Power Automate integration patterns. You will apply these skills to two HR agent topics.

This module builds the prompt design and management layer on top of the conversation flow architecture developed earlier in the short course—covering the three core prompt engineering techniques for enterprise agents, enterprise prompt library architecture, and an off-platform lab that produces a full Copilot Studio conversation architecture and prompt library.

This module develops the MCP connector and Computer Use design skills needed to connect Copilot Studio agents to enterprise systems and legacy applications that fall outside native connector coverage. Learners design MCP connectors that expose external tools and data sources as agent-accessible capabilities, and design Computer Use automations for legacy systems without APIs–applying the decision framework that determines which approach is appropriate for each integration scenario.

This module covers A2A interoperability patterns—including how Copilot Studio agents participate in A2A inter-agent communication, the interoperability patterns for handoffs between Copilot Studio and Foundry agents, and how to validate A2A sequence diagrams for protocol gaps and security vulnerabilities. Learners apply these skills in the off-platform lab that produces the full procurement agent extensibility architecture.

This module develops the M365 Copilot agent design and surface alignment skills needed to architect AI experiences across Microsoft 365 collaboration surfaces. Learners design agent personas and knowledge source mappings for M365 Copilot, apply the extensibility model decision framework to determine when declarative agents are appropriate vs. Copilot Studio agents, and produce a surface alignment design for a provided enterprise sales scenario.

This module builds the Copilot for Sales and Service configuration layer, covering CRM and service data integration, persona and knowledge source specification, and business scenario mapping. Learners apply these skills to propose and configure M365 agents appropriate for the provided enterprise sales and service scenarios through a mix of structured practice and AI-graded architectural reasoning.

Learners produce an integrated Copilot Studio and Microsoft 365 Copilot agent architecture document for a provided enterprise scenario—synthesizing topic architecture, conversation flow design, extensibility design, MCP connectivity, Computer Use automation, A2A orchestration, and Microsoft 365 Copilot surface alignment into a single coherent architecture deliverable that mirrors the platform design artifacts a practicing Copilot Architect produces for enterprise clients.