What an Agent Is and What It Is Not
By the end of this module you can say precisely what makes something an agent rather than a scripted workflow, and recognise how often the scripted workflow is the better answer.
Introduction to Agentic AI and Workflow Automation
An agentic workflow design with trigger, state, tools, approval points, logs, evaluation cases, and a do-not-automate boundary.
Tasks, Goals, Instructions, and Stopping Rules
By the end of this module you can write a task with a goal, instructions and - the part everyone forgets - a stopping rule that ends the run whether or not it succeeded.
Tools, Permissions, and Safe Action Boundaries
By the end of this module you can decide what an agent is allowed to do, and design each permission so the worst-case action is one you could undo.
State, Memory, Context, and Handoff
By the end of this module you can decide what an agent carries between steps and what it hands off, and recognise that most 'memory' problems are really context problems.
Planning, Routing, and Deterministic Workflow Steps
By the end of this module you can separate the steps that must be deterministic from the ones that genuinely need judgement, and stop paying a model to do arithmetic.
Human Approval and Escalation
By the end of this module you can place a human approval point where it actually protects something, and design the escalation path for when the agent is uncertain.
Failure Modes: Loops, Over-Action, Hallucinated Tools, and Stale Context
By the end of this module you can name the ways agentic workflows fail - loops, over-action, invented tools, stale context - and put a specific limit against each one.
Evaluating an Agentic Workflow
By the end of this module you can evaluate an agentic workflow on what it did, not just what it produced: which tools it called, in what order, and where it should have stopped.
Designing Workplace Automations Safely
By the end of this module you can design a workplace automation that a compliance reviewer would sign, with a defined scope, an audit trail, and actions that are reversible.
Capstone: Agent Workflow Design and Risk Review
In this capstone you will design an agent workflow end to end and present it with its permission boundaries, its stopping rules, its approval points, and an honest risk review.
First-release boundary
This is a text-first course shell. Leonardo/image creation, image QC, dependency-heavy runtime QA, external review, and learner pilot evidence remain separate later quality steps.