Building AI Agents and Autonomous Workflow Systems

Craft autonomous agentic AI systems

AI workflows can follow predefined rules, but agentic systems go further by reasoning about goals, selecting tools, and deciding what to do next.

On this ExpertTrack you’ll learn how to design and evaluate autonomous AI systems built around single and multiple agents.

Start with the foundations of agent architecture, including the agent loop, tool use, and different forms of memory.

You’ll compare reactive, goal-directed, and memory-augmented agents and explore how system prompts and structured outputs shape agent behaviour.

Build single-agent systems

Design agents that can interpret instructions, call tools, maintain useful context, and return structured results.

You’ll explore tool configuration and methods for evaluating whether an agent has completed its assigned task successfully.

Coordinate multi-agent workflows

Progress to orchestrator and worker architectures, task decomposition, role assignment, and context passing between agents.

Learn how sequential and parallel execution patterns support more complex workflows while shared memory and state management help agents coordinate.

Evaluate AI agent performance

Explore how to assess task completion, handoff quality, looping behaviour, and common failure modes across an agent system.

By the end, you’ll understand how to design, coordinate, and evaluate agentic AI systems for complex automated tasks.

This ExpertTrack is for automation engineers, AI developers, solutions architects, and technology professionals with workflow experience who want to design single- and multi-agent systems that reason, use tools, and act autonomously.

No prior coding experience is required. You will need a suitable computer, a stable internet connection, and access to the no-code AI agent platforms used in the practical activities. This course uses tools such as n8n, Google API credentials, and Flowise to build, test, and connect AI agents.

This ExpertTrack is for automation engineers, AI developers, solutions architects, and technology professionals with workflow experience who want to design single- and multi-agent systems that reason, use tools, and act autonomously.

• Describe the core concepts, agent types, and business use cases of agentic AI.
• Practice building tool-enabled AI agents and visual agent workflows using n8n and Flowise.
• Demonstrate how memory, context, state, and RAG improve agent behaviour and knowledge grounding.
• Design multi-step agentic workflows that include review, scoring, approval, publishing, webhooks, and event triggers.
• Investigate multi-agent coordination patterns, shared memory, A2A communication, handoffs, troubleshooting, and fallback handling.