Build your first testable AI workflow
Move from an attractive idea to a bounded task, a one-page brief, and a ten-case pilot you can review with another person before investing further.
BEFORE LESSON ONE
Bring real material so the course produces a real result.
- 01 · WHO IT IS FOR
- For anyone turning a repeated knowledge task into a first AI-assisted workflow.
- 02 · WHAT TO BRING
- Bring one real task and several examples of its input and output.
- 03 · HOW TO FINISH
- Another person can run the pilot and reach the same pass/fail decision.
WHEN THIS COURSE FITS
Start here when any of these situations is true.
You do not need a level label. Enter when the work shows one of these signals, bring real material, and use course acceptance to decide when you are done.
- 01✓
You are handing repeated knowledge work to AI for the first time
Real examples and a human workflow exist, but the scope still says ‘automate the whole process.’
- 02✓
The team debates tools without defining success
People disagree about acceptable output, prohibited actions, and who takes over.
- 03✓
A build is proposed without low-cost evidence
You need a small pilot to learn value, risk, and the first failure worth fixing.
COURSE MILESTONES
Every stage leaves a change someone else can review.
These milestones are not another reading list. They show whether the lessons form one piece of work: real material at the start, connected artifacts in the middle, and evidence another person can reproduce at the finish.
- 01 BEFORE↓
Sample one real, repeated knowledge task.
Collect the current workflow, several routine inputs, two difficult inputs, and the way a person judges and hands off the work today. Do not choose a model or treat an entire role as one task.
Starting evidence: sanitized real examples and the current handling steps. - 02 MIDPOINT↓
Turn ‘use AI’ into a bounded working brief.
Complete the fit decision, task brief, and data boundary in sequence. A colleague should be able to name the exact AI action, visible fields, and conditions that stop the workflow.
Midpoint evidence: matrix, brief, and data boundary use the same task and inputs. - 03 FINISH✓
Ask a second person to rerun the ten-case pilot.
Include routine, difficult, mixed-intent, and prohibited inputs. Give the reviewer the materials without explaining your intent and see whether they reach the same pass, revise, or stop judgment.
Completion evidence: the reviewer reproduces the pilot and identifies the next failure to fix.
- 01✓LESSON 1
Should this task use AI at all?
Judge a task by variation, verifiability, stakes, data boundaries, and fallback before choosing a model or building a workflow.
LESSON OUTPUT · decision matrixNot completeOpen lesson → - 02✓LESSON 2
Turn an AI idea into a testable task brief
Replace a vague AI feature request with a one-page brief that names the user, input, acceptable output, evidence, boundaries, and fallback.
LESSON OUTPUT · one-page briefNot completeOpen lesson → - 03✓LESSON 3
Draw the data boundary before AI sees the input
Route every field through allow, transform, isolate, or exclude so a useful workflow does not quietly become an uncontrolled data transfer.
LESSON OUTPUT · data boundary mapNot completeOpen lesson → - 04✓LESSON 4
Run a ten-case AI pilot before you scale
Build a deliberately small test set with ordinary, difficult, ambiguous, and unsafe cases so a promising demo becomes evidence.
LESSON OUTPUT · pilot sheetNot completeOpen lesson →
COURSE ACCEPTANCE
Another person can run the pilot and reach the same pass/fail decision.
Reading is not the finish line. Assemble the lesson outputs and ask another person to review them against the standard above.