AI BootCamps (7 Courses)
Course Category
Course Category
Google Rating
Professionals Upskilled
Live Cohorts Delivered
Enterprise Teams Trained
In this 4-hour live Creating Agents in Copilot Studio program, learn to build, test and publish a functional AI agent using Microsoft Copilot Studio. Learn how to design agent behaviour, connect enterprise knowledge, use business tools, automate workflows, add approvals, evaluate responses and securely publish the agent for real users.
Build an agentic workflow with Claude - Understand the agent loop and learn how Claude moves from instructions to planning, tool use, execution and results.
Give Claude project-specific context - Use CLAUDE.md and structured instructions to teach Claude how your project, codebase or workflow works.
Connect Claude to external tools - Use MCP to connect Claude with services, APIs, repositories, databases and other external capabilities.
Equip agents with reusable Skills - Create reusable Skills that package instructions, resources and scripts into capabilities Claude can use when needed.
Orchestrate specialised agents - Use subagents to delegate specialised tasks and coordinate multiple pieces of work from a primary agent.
Test, control and improve agents - Learn how to review agent actions, manage permissions, test outcomes and improve agent reliability before allowing greater autonomy.
Agent fundamentals
Lab 1: Create your first Claude agent
Outcome:
Project context
CLAUDE.md as persistent project instructionsTools and MCP
Anthropic’s current Claude ecosystem supports MCP integrations for connecting Claude Code to external systems, while advanced tool-use capabilities allow agents to discover and work with larger tool libraries.
Lab 2: Connect Claude to external capabilities
CLAUDE.mdOutcome:
Agent capabilities
SKILL.mdAgent Skills are designed as reusable folders containing instructions, resources and, where appropriate, executable code that Claude can load when relevant. Anthropic
Subagents and orchestration
Lab 3: Build a specialised agent workflow
Outcome:
Agent reliability
Claude Code currently supports capabilities such as checkpoints, subagents, hooks and background tasks for longer-running agentic workflows. Anthropic
Lab 4: Test and complete the agent
Outcome:
This 4-hour course is not a certification-preparation program. It provides practical foundations that can support further preparation for relevant Claude certifications.
Claude Code changed how I think about AI-assisted development. The hands-on work with project context, tools and agent workflows made the difference between simply prompting Claude and actually delegating work to it.
The MCP and subagent exercises were particularly valuable. I could see how Claude can move from a coding assistant into an agent that connects to real systems and handles larger workflows.
The course gave us a practical framework for using Claude safely. We learned how to give the agent context, let it execute meaningful work, verify the results and keep humans in control.
You will learn how to build agentic workflows with Claude, provide project context, connect external tools through MCP, create Skills, use subagents and validate agent-driven work.
Yes. This course is designed for technical professionals and requires basic programming, Git and terminal knowledge.
Yes. The entire course is built around hands-on labs, with one AI software engineering agent developed progressively throughout the session.
No. The course introduces Claude’s agent workflow and progressively moves into MCP, Skills and subagents.
No. It is a practical agent-building course. It can provide useful foundations for professionals pursuing relevant Anthropic Claude certifications.