Automation and AI solutions
AI agents for business course
Agents that act on your systems, with clear limits and human oversight.
An AI agent does not just answer: it uses tools, queries systems and takes actions to achieve a goal. It is the next step beyond assistants and the most promising way to automate complex processes, and also the one that demands the most care.
This course teaches you to build agents that work in a real company: well-designed tools, connection to systems through MCP, proven design patterns, human approval before irreversible actions, protection against instruction injection and evaluation with test tasks.
We start with the most important point: knowing when you do not need an agent and a fixed workflow is enough.
What you will learn
- Understand what an agent is and when it beats a fixed workflow
- Give an agent tools and connect it to systems through MCP
- Design multi-agent workflows with human approval points
- Protect agents from misuse and costly mistakes
- Evaluate and monitor agents in production
Syllabus
8 modules, from the fundamentals to the final project.
-
What an agent is
- Model, tools, loop and goal
- Workflows versus autonomous agents
- Realistic business use cases
-
Tools
- Function calling (tool use)
- Designing clear, safe tools
- Model Context Protocol (MCP) for connecting systems
-
Planning and memory
- Reasoning, planning and self-correction
- State, short-term and long-term memory
- Managing context in long tasks
-
Design patterns
- Chaining, routing and parallelisation
- Orchestrator and workers
- Evaluator and optimiser
-
Human in the loop
- Approvals before irreversible actions
- Escalation and handover to a person
-
Security
- Least privilege and sandboxed environments
- Instruction injection through data
- Spending and action limits
-
Evaluation and observability
- Execution traces
- Sets of test tasks
- Success, cost and time metrics
-
Frameworks
- Model providers' SDKs
- LangGraph and other agent frameworks
- Agents in low-code platforms
Tools you will practise with
- Python
- Claude Agent SDK or OpenAI Agents SDK
- LangGraph
- Model Context Protocol (MCP)
- n8n
Hands-on project
You build an agent that completes a multi-step task on a company's simulated systems (for example, handling a customer issue by checking orders and proposing a solution), with tools, human approval for sensitive actions and a set of evaluation tasks.
Requirements
- Python programming.
- Ideally, having taken the LLM application development course or having equivalent experience.
Other courses
-
AI automation with low-code tools course
Automate complete processes by combining AI and visual tools, without coding.
- Duration: 12 hours
- Level: Beginner
-
Copilot Studio and Power Automate course for citizen developers
Build AI agents and automations in Microsoft 365 without coding, and within the rules IT needs.
- Duration: 12 hours
- Level: Intermediate
-
Corporate chatbots course: design and deployment
Design a chatbot that genuinely resolves queries and knows when to hand over to a person.
- Duration: 10–12 hours
- Level: Intermediate
Where it is applied
-
AI process automation
We automate repetitive tasks with AI so your team spends less time copying data, reviewing documents, classifying requests or drafting replies by hand.
-
AI agent for preparing quotes
An AI agent reads each request, checks prices, stock and lead times in your systems and prepares the quote for a person to review and send.
Related articles
-
Use cases
AI agents in your company: what they are, what they can do and how to keep them in check
What an AI agent is, how it differs from a chatbot or an automation, examples in an SME, the risks and how to start with human approval.
5 min read
Want this course for your team?
We adapt the content, length and format to your company.