Building Process-Driven Agents in Microsoft Copilot Studio (Standard Harness)

This course is intended for:

• Power Platform developers and low-code makers, who want to design and deploy AI agents integrated with enterprise data and systems.
• IT professionals and Microsoft 365 administrators, responsible for sharing, securing, and managing agents across the organization.
• Business analysts and process owners, who want to automate repetitive tasks and improve the quality of service for internal users and customers.
• Solution consultants and architects, who design Copilot Studio implementations aligned with governance best practices and ALM.
• People getting started with AI agents, who want a structured introduction to the Copilot Studio platform and its capabilities.

After completing this course, participants will gain the following benefits:

1. A complete “from idea to deployment” workshop – independently building, testing, and publishing a working agent.
2. An agent grounded in enterprise knowledge – grounding responses in data from SharePoint, files, and public websites (grounding / RAG) while respecting user permissions.
3. Business process automation – interactive forms (Adaptive Cards) and agent flows integrating SharePoint and email.
4. Autonomous agents – responding to events (event triggers) and performing actions without user involvement, within safe boundaries.
5. Secure publishing and distribution – making the agent available in Microsoft Teams and Microsoft 365 Copilot in line with governance and DLP policies.
6. Cost and licensing control – understanding the Copilot Credits model (pay-as-you-go / capacity packs) and estimating costs before scaling.
7. Governance and application lifecycle management (ALM) – packaging the agent in solutions, publishers, security roles, and promotion DEV → TEST → PROD.

In practice, this means that you will leave the course with a ready-to-use, working agent and the knowledge of how to apply this pattern to real processes in your organization – faster, cheaper, and more securely than building a solution from scratch.

Module 1 – Introduction to Agents and Copilot Studio

• Chatbot vs. AI assistant vs. agent – key differences and business value
  • The LLM as the agent’s “brain”, RAG (grounding), and generative vs. classic orchestration
  • Copilot Studio within the Microsoft AI platform; choosing a harness: GitHub Copilot, standard, Copilot Chat
• Standard harness: new vs. classic authoring experience
  • Agent building blocks: Knowledge, Tools, Topics, Instructions; environments and the model

Module 2 – Power Platform Solutions

• Solutions as containers that package the agent and its components (ALM)
  • Managed vs. unmanaged solutions – when to use each
• Publishers and prefixes; security roles (Environment Maker, System Customizer, System Administrator)
  • Solution lifecycle DEV → TEST → PROD; solution explorer in Copilot Studio
  • Lab: creating a publisher and a custom unmanaged solution for the agent

Module 3 – Building Agents on the Standard Harness

• Creating an agent from a natural language description; defining instructions
  • Grounding the agent: SharePoint, uploaded files, public websites; controlling general knowledge
• Topics, trigger phrases, and conversation nodes; variables and entities
  • Testing in the Test your agent pane and iteratively improving response quality
  • Labs: building an agent through conversation and adding a custom topic with a trigger and nodes

Module 4 – Designing Interactive and Automated Experiences

• Adaptive Cards – interactive forms built in JSON and dynamically with Power Fx
  • Agent flows – deterministic automation that the agent calls as a tool
  • Integration with SharePoint and email; Power Fx filters and expressions
  • Labs: a device selection card and a flow that handles the request and sends an email notification

Module 5 – Building Autonomous Agents

• Conversational vs. autonomous agents; the components of an autonomous agent
  • Event triggers – responding to events in SharePoint, Teams, and Outlook; payload anatomy
  • Authentication, security, and human-in-the-loop (guardrails, least privilege)
• Effective agent instructions: role, scope, context, tone, error handling, and escalation
  • Lab: automatically responding to a new SharePoint request with an email confirmation

Module 6 – Testing, Publishing, and Distributing Agents

• Testing before publishing; the difference between testing and publishing (and republishing)
  • Distribution channels: Microsoft Teams and Microsoft 365 Copilot; governance, DLP, and approval in the Teams admin center
  • Licensing and costs: Copilot Credits, pay-as-you-go and capacity packs; estimating consumption
  • Lab: publishing the agent to Teams and Microsoft 365 Copilot and sharing it across the organization

• Basic knowledge of Microsoft 365 services (Teams, SharePoint, Outlook) at user level.

• General understanding of cloud services and of artificial intelligence and large language model (LLM) concepts – an advantage, not a requirement.

• No programming experience is required – the course follows a low-code / no-code approach.

• Prior familiarity with Microsoft Power Platform (environments, solutions) is recommended and makes getting started easier.

•  manual in electronic form available on the platform: https://www.altkomakademia.pl/
• access to Altkom Akademia's student portal

Training method:

• Lecture: 30%

• Workshops: 30%

• Labs:40%

• Training: English

• Materials: English