AI & Generative AI · AI Agents & Automation
Design AI Agents and Automated Workflows
Learn how AI agents plan, use tools and complete multi-step tasks, and how to build automated workflows with the evaluation, monitoring and guardrails they need to run safely.
- Agent architecture and planning
- Tool calling and integrations
- Workflow orchestration
- Evaluation, monitoring and guardrails
Course overview
About This Course
AI Agents & Automation explains how to build systems where AI models carry out multi-step tasks: deciding what to do next, calling tools, and working through a workflow towards a goal.
Agents and AI-driven automation can take on repetitive, multi-step work, but they also make mistakes and act on real systems. Building them well means designing clear workflows, limiting what they can do, and measuring how they behave.
You will learn common agent architectures, how to define and secure tools, how to orchestrate multi-step workflows, how to keep a human in the loop for sensitive steps, and how to evaluate and monitor agents.
This course builds on Generative AI Development in our AI & Generative AI learning paths. AI Security covers the security of agents and tool access in more depth.
Who it is for
Who Should Take This Course?
-
Developers building with LLMs
You have used LLM APIs and want to build systems that take multi-step actions.
-
AI engineers
You want to design agents that are reliable enough for real use.
-
Automation specialists
You automate business processes and want to add AI-driven steps.
-
Technical leads
You need to judge where agents add value and how to control them.
Learning outcomes
What You Will Learn
By working through the course, you will build towards being able to:
Explain common agent architectures and when an agent is, or is not, the right approach.
Define tools with clear inputs, outputs and permissions.
Orchestrate multi-step workflows that combine AI steps with regular code.
Manage state, memory and context across steps.
Add human approval and other guardrails for sensitive actions.
Evaluate agent behaviour with task-based tests.
Monitor agents in operation and investigate failures.
Curriculum
Course Curriculum
The course is organised into modules. Session order and depth may vary by training arrangement.
-
Module 1: Agent fundamentals
What makes a system an agent, and when to use one.
Topics covered include
Practical activity: Compare an agent and a fixed workflow for the same sample task.
-
Module 2: Tool calling
Give models safe, well-defined ways to act.
Topics covered include
Practical activity: Define a set of tools for a sample support-desk assistant.
-
Module 3: Workflow orchestration
Combine AI steps and regular code into dependable flows.
Topics covered include
Practical activity: Orchestrate a multi-step document-processing workflow.
-
Module 4: Multi-step AI workflows
Build agents that work through tasks towards a goal.
Topics covered include
Practical activity: Build an agent that researches a question using sample tools and summarises the result.
-
Module 5: Guardrails and human oversight
Keep agents within safe limits.
Topics covered include
Practical activity: Add approval steps and limits to the sample agent.
-
Module 6: Evaluation and monitoring
Measure how agents behave and keep them reliable.
Topics covered include
Practical activity: Evaluate the sample agent on a set of tasks and fix the most common failure.
Practical learning
Hands-On Learning
Practical work builds and improves a sample agent. Examples of activities include:
Activities use sample tools and data in a practice environment. Some use AI model APIs; how access is arranged will be explained when you enquire.
- Designing tools with clear permissions
- Orchestrating multi-step workflows
- Adding guardrails and human approval
- Evaluating and tracing agent behaviour
Before you start
Prerequisites and Learning Requirements
-
Recommended background
Experience building with LLM APIs is recommended. Generative AI Development covers this.
-
Technical prerequisites
Programming experience, typically in Python, and familiarity with APIs and JSON.
-
Software and tools
A computer with a reliable internet connection for live online sessions. Python and a code editor are used for practice. Any additional software or accounts for practical activities will be confirmed when you enquire.
-
Prior experience
Not intended for beginners in programming or AI.
Learning Format
IT Champion Hub offers live, instructor-led online training and workshops. Formats, duration and schedules vary by course and are not the same for every program.
- Live online sessions led by an instructor
- Practical activities alongside explanations
- Upcoming batches and schedules shared on enquiry
Enquire to find out the formats and upcoming batches available for this course. Enquire about this course
FAQ
Frequently Asked Questions
Do I need to take Generative AI Development first?
It is recommended. You should already be comfortable calling LLM APIs, writing prompts and handling structured output.
Which agent frameworks are covered?
The course focuses on agent design principles that apply across frameworks. Ask us which tools are used in your training arrangement.
How do you keep agents safe?
The course covers least-privilege tools, input and output checks, human approval for sensitive actions and monitoring. AI Security goes deeper into securing agents.
Are there practical exercises?
Yes. You will build, test and improve a sample agent throughout the course.
How can I enquire about upcoming training?
Select Enquire About This Course. The enquiry form will note that you are asking about AI Agents & Automation, and we'll get back to you with upcoming batches and available schedules.
Ready to Explore This Learning Path?
Have questions about the curriculum, prerequisites, or training arrangements? Get in touch with IT Champion Hub to discuss your learning goals.