Overview
This module explores how AI agents can work together with tools, triggers, external applications and automated workflows to handle more complex business activities.
Building on the agent creation concepts covered in the previous module, learners will take a deeper look at Microsoft Copilot Studio and understand how to design agents that can respond to specific conditions, perform actions and connect with other business systems.
The module covers conversational topics, variables, triggers, contextual instructions, Power Automate integration and Model Context Protocol (MCP). Learners will also explore how multiple agents can be connected to create a coordinated AI agentic framework.
By the end of this module, learners will understand how to move beyond individual AI assistants and design connected agent workflows that support business processes.
Tools Covered: Microsoft Copilot Studio – Advanced Agent Configuration | Topics | Variables | Triggers | Tools | Knowledge Sources | Power Automate | Model Context Protocol (MCP)
What Will Be Covered
10.1 Create Focused Topics with Variables, Triggers & Conversational Flow
Learn how to structure conversations in Copilot Studio so agents can understand user requests and respond through defined conversational paths.
Topics include:
Understanding topics and their role in managing agent conversations.
Creating focused topics for specific business requirements.
Configuring triggers to identify when a topic should start.
Using variables to capture and store information during conversations.
Designing conversational flows with questions, responses and conditions.
Managing different conversation paths based on user inputs.
Connecting related topics to create a smooth interaction.
Practical example: Create an employee support agent that identifies whether a user needs help with leave requests, payroll or IT support and guides them through the appropriate conversation.
10.2 Ground Chain-of-Thought Prompts with Tools
Explore how structured instructions and connected tools can help agents handle tasks that require multiple steps.
Topics include:
Understanding multi-step reasoning instructions for AI agents.
Structuring prompts to break complex tasks into manageable steps.
Connecting tools to help agents retrieve information or perform actions.
Using relevant business context when an agent works with tools.
Guiding agents to select suitable tools based on the user's request.
Reviewing tool outputs before generating a final response.
Understanding the importance of human review for important business decisions.
Practical example: Configure an agent to review a business request, retrieve relevant information from a connected source, use an available tool and prepare a response based on the retrieved results.
10.3 Trigger Agent Workflows Based on a Condition and Trigger External Workflows from Agents Using Power Automate
Understand how agents can initiate actions and automate business processes based on specific conditions.
Topics include:
Understanding event-based and condition-based triggers.
Identifying business events that can initiate an agent workflow.
Configuring conditions to control when an action should take place.
Connecting Copilot Studio agents with Power Automate flows.
Triggering external workflows from an agent conversation.
Passing information between agents and automated workflows.
Managing workflow responses and handling exceptions.
Reviewing the results of automated actions.
Practical example: Create an approval support workflow where an agent collects an employee's request and triggers a Power Automate flow to notify the relevant manager when the request meets predefined conditions.
10.4 Tag Triggers, Topics, Tools & Knowledge Sources with Contextual Instructions
Learn how to organise an agent's components and provide instructions that help it use the correct topics, tools and knowledge sources.
Topics include:
Understanding the relationship between instructions, triggers, topics and tools.
Organising agent components based on their purpose.
Providing contextual instructions for different topics and business functions.
Defining when an agent should use a particular tool or knowledge source.
Connecting relevant knowledge sources to specific business requirements.
Guiding agents to select appropriate actions based on conversation context.
Reviewing how instructions influence agent behaviour and responses.
Practical example: Configure a customer support agent with separate instructions for billing, technical support and service enquiries, ensuring that each request is handled using the relevant topic, tool and knowledge source.
10.5 Ground Machine Learning Models with AI Instructional Prompts
Explore how AI instructions can guide agents when working with machine learning models and their outputs.
Topics include:
Understanding how machine learning model outputs can be used within agent workflows.
Preparing contextual instructions for interpreting model predictions.
Connecting relevant business information with model outputs.
Guiding agents to explain predictions in a business-friendly format.
Using instructions to define how forecasting and prediction results should be presented.
Understanding the limitations of model outputs and the importance of validation.
Using human review when predictions influence important business decisions.
Practical example: Configure an agent to interpret project delay predictions and explain which projects may require management attention based on the model output and relevant project information.
10.6 Create & Integrate MCP Servers & Protocols for Internal & External App Actions and Data Retrieval
Understand how Model Context Protocol (MCP) can support connections between AI agents and external tools, applications and data sources.
Topics include:
Introduction to MCP and its role in AI agent integration.
Understanding MCP servers, clients and communication protocols.
Exploring how MCP enables agents to access tools and retrieve information.
Understanding connections with internal business applications and external services.
Exploring how agents can use MCP-enabled tools to perform defined actions.
Understanding data retrieval and tool execution through MCP connections.
Identifying suitable MCP integration opportunities for business requirements.
Reviewing access permissions and security considerations when connecting external systems.
Practical example: Explore how an AI agent can connect to an external business service through an MCP server to retrieve relevant information or perform an approved action.
10.7 Build Connections Between AI Agents & Develop a Comprehensive AI Agentic Framework
Learn how multiple AI agents can be connected to handle different responsibilities within a larger business process.
Topics include:
Understanding multi-agent systems and their role in business workflows.
Identifying separate responsibilities for specialised agents.
Connecting agents based on their tasks and capabilities.
Coordinating information exchange between agents.
Using tools, triggers, knowledge sources and workflows across connected agents.
Designing a process where one agent's output becomes the input for another agent.
Reviewing workflow execution and identifying areas for improvement.
Understanding the importance of monitoring, permissions and human oversight.
Practical example: Design a business workflow where one agent collects a customer request, another retrieves relevant account information and a third prepares a response or initiates an approved follow-up action.
Skills You Will Gain
By completing this module, learners will develop the ability to:
Create focused conversational topics in Copilot Studio.
Configure variables, triggers and conversational flows.
Design structured instructions for multi-step agent tasks.
Connect agents with tools and relevant business information.
Build condition-based agent workflows.
Integrate Copilot Studio with Power Automate.
Configure contextual instructions for topics, tools and knowledge sources.
Guide agents in interpreting machine learning outputs.
Understand MCP servers, clients and protocols.
Explore connections between AI agents and internal or external applications.
Design workflows involving multiple connected AI agents.
Understand the importance of permissions, testing and human oversight in agentic systems.
Learning Outcomes
After completing this module, learners will understand how to design connected AI agent workflows that can support business operations and reduce repetitive manual activities.
More structured business interactions: Create agents that guide employees and customers through defined conversations, collect relevant information and direct requests to the appropriate process.
Reduced manual coordination: Use triggers and automated workflows to initiate follow-up actions, notifications and approvals without requiring employees to manage every step manually.
Connected business applications: Understand how agents can work with internal and external systems through tools, integrations and MCP connections.
Improved process execution: Design workflows where agents handle different responsibilities, pass information between steps and support the completion of business tasks.
Better use of business information: Configure agents to use relevant knowledge sources and contextual instructions when responding to requests or performing actions.
More effective decision support: Guide agents in interpreting machine learning predictions and presenting results in a format that business users can understand.
Improved workflow consistency: Use defined topics, conditions, tools and instructions to make business processes more structured and repeatable.
Scalable AI adoption: Understand how specialised agents can be connected to support larger business processes across departments rather than relying only on individual assistants.
AB-731 Exam Alignment
Learning area | Relevance |
|---|---|
Understanding AI agents | Explores how agents can respond to instructions, use tools and support business activities. |
Microsoft AI solutions | Provides practical exposure to Copilot Studio and its agent-building capabilities. |
Business process automation | Demonstrates how agents and Power Automate can support automated business workflows. |
AI integration | Introduces connections between agents, knowledge sources, tools and external applications through MCP. |
AI adoption and business value | Shows how connected agents can reduce repetitive work and improve process execution. |
Responsible AI implementation | Highlights the importance of permissions, testing, monitoring and human oversight. |
Still need help?
Contact us