Building AI Agents Without Code
Learners ready to turn a bounded business process into a no-code agent with clear human controls.
What you will be able to do
- Define an agent's job, inputs, decision rules, and hand-offs.
- Build a no-code agent for a bounded business use case.
- Test edge cases and diagnose failure modes.
- Configure human approval for consequential actions.
- Present an agent performance and risk log.
How this is assessed
Working agent — lead qualifier, inbox triage or report builderSubmit a live no-code workflow with your written test cases and operating notes. Include the risk log and enough access or recording for a reviewer to see it actually run.
| Grade | What earns it |
|---|---|
| Pass | A functioning agent that does the job described, end to end. |
| Merit | Reliable edge-case handling, evidenced by written tests and a changelog. |
| Distinction | Robust controls and fallbacks, a named owner, and a business case a manager could act on. |
Before you submit
- The agent's job is stated in one sentence.
- Written test cases cover normal, ambiguous and failure inputs, with expected outcomes.
- A human approval sits in front of every consequential action.
- For merit: edge cases are handled reliably and the changelog shows before-and-after evidence.
- For distinction: controls, fallbacks, a named owner and a clear business case.
Handing it in
Marked by a no-code automation builder with deployed AI workflows in commercial or operations teams. You will need an account to submit, so that the mark has somewhere to land. Sign in or create one. The modules are open either way.
Modules
8 modules, self-paced. Nothing is timed and nothing is scored.
- 1Agent scopeChoose a repeatable, bounded process.
- 2Decision designDefine inputs, rules, and exceptions.
- 3Knowledge setupStructure instructions and reference material.
- 4Workflow buildConnect trigger, agent, and destination.
- 5Testing labRun normal, ambiguous, and failure cases.
- 6ControlsAdd approvals, logging, and fallback paths.
- 7IterationImprove from test evidence, not from vibes.
- 8The risk logWhat it does well, what it cannot do, who owns it.
This course stacks
Can take a real business process, give an AI the context it needs, and automate it end to end without losing accountability.
A credential is issued when all of its artifacts have passed. It records the courses, the bands and the dates, nothing else, and no score.