Agentic AI Workflow & Execution Design
Move from a single AI task to a repeatable agentic workflow: learn task decomposition, sequential / parallel / routing patterns, decision points, the agent loop, execution, verification and human-in-the-loop, decide which steps stay fixed and which the agent may decide, and complete your own Agentic AI Workflow Canvas.
Course storyline: from a task to verified execution
The whole course follows one line: decompose, sequence, route, decide, execute, verify.
- 1. Decompose
What tasks make up this job?
- 2. Sequence
What must happen in order?
- 3. Route
When should the workflow branch?
- 4. Decide
Which steps require agent judgment?
- 5. Execute
What action should happen next?
- 6. Verify
How do we know the result is acceptable?
Your next step
AI 104 · Module 1: From Task to Agentic Workflow
Curriculum
- Module 1From Task to Agentic WorkflowWhat is the difference between a single prompt, a fixed workflow and an agent?
- Module 2Task Decomposition & Execution StepsHow should a complex job be broken into steps an agent can execute?
- Module 3Sequential, Parallel & RoutingShould all work run in a fixed order?
- Module 4Decision Points & Agent LoopWhich steps can be fixed in advance, and which need the agent to decide?
- Module 5Execution, Verification & Human-in-the-LoopAfter the agent takes the next step, how do we know we can continue?
- Module 6Agentic AI Workflow CanvasFrom design ideas to a workflow ready for the next stage
Progress is stored in this browser and does not identify you. It is uploaded only if you join progress sync below.
Sync my progressJoin a session to keep progress across devices and share it with your instructor
Practice and assessment
- Activity 1: Prompt, Workflow or Agent?Judge whether a situation calls for a single prompt, a fixed workflow, an agentic workflow or an agent.Not done yet
- Activity 2: Task DecompositionBreak a real job into executable tasks, inputs and outputs.Not done yet
- Activity 3: Workflow PatternChoose sequential, parallel, routing or loop for a scenario.Not done yet
- Activity 4: Fixed or Agentic?Tell which steps should be fixed, which need an agent decision and which need a human decision.Not done yet
- Activity 5: Execution & HITLAdd verification, retry, fallback and human-in-the-loop to a workflow.Not done yet
- AI 104 assessmentTwo levels: knowledge checks and scenario decisions.Not done yet
- Agentic AI Workflow CanvasBring everything together in an Agentic AI Workflow Canvas you can take back to work or a project.Not done yet
About this course
Outcomes
- Distinguish Prompt, Chain, Workflow, Agentic Workflow and Agent
- Decompose complex work into executable tasks and subtasks
- Choose workflow patterns such as sequential, parallel and routing
- Identify fixed, conditional, agent and human decision points
- Add execution, verification and human-in-the-loop to a workflow
- Complete an Agentic AI Workflow Canvas
Prerequisites
- AI 103 is recommended, or a basic grasp of context, memory, knowledge and agent skills
- You know AI output needs verification and high-risk actions must not be left entirely to AI
- No programming background required
Who it is for
- Staff and faculty who finished AI 103 and want to combine AI capabilities into a repeatable workflow
- Coordinators and managers who plan AI processes, automation or digital transformation
- Students and researchers who want to understand workflows, agent decisions and human-in-the-loop
- Learners preparing to study MCP, planning, multi-agent systems or agent harnesses next
We suggest finishing AI 103 first. This is a suggestion only and never blocks you from starting.
Certificate
Earns a course completion certificate (L1); a capability certificate (L2) is also available through additional artifact review.
Requirements to earn it
- Complete all 6 course modules
- Pass the post-test assessment at 80% or higher
- Submit a project artifact and pass review
Course feedback
It takes 2 to 3 minutes and is used to improve the course.