TutorialsAssignment: Write your own SOP card, then turn it into a workbench
Three levels merged into one assignment — run a scenario to get SOP v2, apply it to a repeated task in your own role, then build a first-version workbench. All three required.
Objective
Connect the three lessons into one deliverable you can actually take away:
- Run one scenario: pick daily report, project retrospective or post breakdown, and run the class SOP on your own material;
- Apply it to your own work: pick a task you repeat every week, fill in the AI SOP card, and verify it with a second set of real material;
- Build a first-version workbench: fill in the workbench blueprint, choose 2–4 areas, and produce a requirements table or a single-page HTML prototype.
What earns marks is not how polished the output looks, but that every piece shows traces of a real test run.
Requirements
Submit exactly three things: an SOP v1-to-v2 comparison, a filled-in AI SOP card, and a workbench blueprint (or prototype). All three required — missing one means the assignment is not complete.
- Use real materials: demo data and sample posts do not count. Use material from your own work.
- v1 to v2 required: a single version is not complete — it must show which specific error you caught and which rule you added.
- Format: one document, or paste directly into the submission area; within two screens, no long screenshots.
- Deadline: within 48 hours of this lesson ending, based on the timestamp shown in the submission area.
Check the three red lines first: take facts, numbers, dates and names back to the source material; desensitize or get permission for demo materials; AI output is a draft — publication and scheduling commitments are confirmed by a human.
Part 1 · Run one scenario (SOP v1 to v2)
Pick one of the three scenarios and run it on your own material.
| Item | Your answer |
|---|---|
| Which scenario I chose (daily report / retrospective / post breakdown) | |
| What real material I used (one sentence, desensitized) | |
| The prompt I sent (copy it, do not polish it) | |
| The specific error I caught in the v1 output | |
| The one rule I added because of that error | |
| What second set of material I verified v2 with, and the result |
Do not write the specific error as not good enough or a bit empty. Write which field was wrong, what it became, and what it should have been — for example, it marked login page copy awaiting confirmation as done, or it invented a launch date.
This box is where your assignment shows its level.
Part 2 · Apply it to your own work (AI SOP card)
Pick a task you repeat every week — lesson prep, writing, reporting, client follow-up, data cleanup all qualify.
Note how this differs from Part 1: there you reproduced a class scenario; here you write a new SOP card for a different task. Changing the variable is how you find out whether the method really landed.
My AI SOP card
| Element | Your answer |
|---|---|
| Task and trigger (what makes you start doing it) | |
| Role, goal and boundaries | |
| Inputs and how gaps are handled | |
| Steps (3–6) | |
| Decisions and no-go zones (what must have evidence, what is forbidden) | |
| Output format and one acceptable example | |
| Human review and completion criteria | |
| Error found in the v1 test run and the v2 change |
| Follow-up | Your answer |
|---|---|
| What second set of real material I used | |
| Did the SOP hold up on the second set? Anything that still needs changing |
Part 3 · Build a first-version workbench (blueprint)
Fill in the blueprint before deciding whether to generate a prototype. Do not overload the areas — 2 to 4 is enough for v1.
My workbench blueprint
| Stage | Where the input comes from | What I do | Verifiable output | Where the result goes |
|---|---|---|---|---|
| Goal | ||||
| Project | ||||
| Action | ||||
| Today's execution | ||||
| Review |
| Follow-up | Your answer |
|---|---|
| Which areas I chose for v1 (2–4) | |
| Which stages I deleted, and why they are not needed | |
| What the pinning rule for today's tasks is | |
| What the task completion criteria are | |
| Whether I delivered a requirements table or a single-page HTML prototype | |
| One adjustment I recorded after trying it (write none if there was none, but say why) |
You do not need to keep every stage. If a Gantt chart does not fit your tasks, replace it with a project list — the loop is the skeleton and the interface is replaceable. Deleting a stage is itself a judgment; explaining why is worth more than keeping everything.
Grading
| Level | Standard |
|---|---|
| ⭐⭐⭐ Excellent | All three submitted; the error in Part 1 is specific to a field and the added rule is directly reusable; the SOP card uses a real task from your own role and was verified on a second set of material; the blueprint states a clear pinning rule, completion criteria and one genuine adjustment record |
| ⭐⭐ Good | All three submitted; a complete v1-to-v2 process; the SOP card covers all six elements; the blueprint loop is readable |
| ⭐ Pass | All three submitted; one scenario run through; the SOP card and blueprint are basically complete |
The two most common reasons for missing ⭐⭐⭐ are writing the specific error as a feeling (not good enough, a bit empty) and filling in all five stages of the blueprint without being able to say what comes first today.
Submission
- Upload a screenshot of each finished assignment to the website for submission.
- Pick the piece you are happiest with: post it in the course group so everyone can see your workflow.
One last reminder: what you take home today is not a universal prompt but a method — write your experience into an SOP, verify it with real work, then put repeated workflows into a workbench. Get one thing running first, and the workbench will truly belong to you.