Omar, the owner of Cedar & Co., a local brand that buys from Delta Garments on a hybrid basis, a confirmed base quantity each season plus a forecast top-up decided closer to the time, uses an AI tool for admin work most days: drafting shop communications, summarising sales notes, and preparing short buying updates. Across the last four lessons, Omar has learned to write a clear instruction, review a draft for substance rather than just tone, decide when an AI first draft is worth asking for, and keep a human check on anything that leaves his own team. Each habit is useful alone. Used together, in order, they form one repeatable checklist that works whether the task is a five-minute internal note or a document reaching a factory outside his own company.
This lesson does not add a new rule. It puts the four habits into a single sequence, then applies that sequence to one more real task: a hybrid buying update to Delta Garments that mixes a confirmed order with a forecast figure.
You need: the habits from Lessons 2.1 to 2.4: a clear instruction, a substance check, a task-by-task decision, and a hand-off check.
By the end of this lesson you can:
Not covered here: which specific AI product to use, or a general claim about how much time the checklist saves on every task. The checklist is a repeatable habit, not a guaranteed shortcut.
| Term | Meaning |
|---|---|
| AI-use checklist | The ordered set of checks a person applies to any AI-assisted task, from instruction to hand-off. |
| Confirmed order portion | The part of a hybrid order that is already agreed and fixed. |
| Forecast order portion | The part of a hybrid order that is planned but not yet fixed, subject to change. |
| Instruction | The written request you give an AI tool, stating the task, the facts and the required format. |
| Human check | A person's own review of an AI-assisted output against known facts before it is used or shared. |
| Hand-off point | The moment an AI-assisted output leaves the control of the person or team that produced it. |
Each earlier lesson taught one habit in isolation so it could be learned properly. In real daily work, the four habits happen in one sequence on the same task: first decide whether an AI first draft is worth asking for at all, then write a clear instruction if so, then check the draft for substance, then apply a final human check at the point the output leaves your own team. Skipping the order does not save time; it just moves the risk from one habit to a gap between two of them.
| What can differ | The shapes you will meet | Why it matters to you |
|---|---|---|
| Order type | An export shipment note (Order A) and a domestic buying update (Cedar & Co.) both go through the same four steps, with different facts inside them. | The checklist does not need a separate version for export versus domestic work. |
| Business model | An MTO task has one confirmed set of facts; an MTS task repeats the same checklist on a schedule; a hybrid order, like Cedar & Co.'s confirmed base plus forecast top-up, needs the confirmed and forecast facts labelled separately at both the instruction step and the check step. | A hybrid order is the shape most likely to blur a forecast into a stated fact if the checklist is applied loosely. |
| Who receives the output | An internal note only needs the check step to be quick; a document reaching Delta Garments or another company outside Cedar & Co. needs the hand-off step applied carefully, every time. | The four steps do not change, but how much attention each one needs changes with the audience. |
| Task size | A five-line internal note and a full quotation cover letter both need all four steps; the time spent on each step scales with the task, not the number of steps. | A short task is not an excuse to skip the decide or check steps; it just makes each one quick. |
| Product family | The facts inside the checklist change (a hosiery reorder's dozen-pair quantities, a bedding set's fabric approval), but the four-step sequence applies to every product family in the master cast. | Learning the checklist once means it works on any task Omar or a colleague meets. |
Cedar & Co. buys from Delta Garments on a hybrid basis: a confirmed base quantity agreed each season, plus a forecast top-up decided closer to delivery. Omar needs to send Delta Garments a short update: the confirmed base quantity for the coming season, and Cedar & Co.'s current forecast for the top-up, clearly marked as not yet final.
Decide: the facts are mostly settled (the confirmed base quantity is fixed; the forecast is a known estimate), and checking two clearly labelled numbers is quick, so Omar asks an AI tool for a first draft.
Instruct: Omar states the task (a short buying update to Delta Garments), gives the facts (confirmed base quantity, forecast top-up figure, and that the forecast is not final), and sets the format (two short paragraphs, one for each portion, clearly separated).
Check: the AI-drafted update reads well, but one sentence merges the confirmed and forecast figures into a single stated quantity. Omar corrects it, restating the confirmed figure and the forecast figure as two separate numbers with their own labels.
Hand off: before sending, Omar rereads the corrected update once as Delta Garments will read it, confirms both figures against Cedar & Co.'s own real records, and only then sends it.
The worked numberOmar tracks which steps mattered for this particular task.
All four steps were needed on this one task, and the step most people would be tempted to skip, the check, was the step that caught the only real error. This matches the pattern across the whole course: a shorter checklist would have let the merged figure reach Delta Garments as if it were a single confirmed quantity.
Once a person has successfully used the four steps a few times, it is tempting to stop deliberately stepping through them and rely on general good habits instead. A common mistake is skipping the explicit decide or check step on a task that looks routine, precisely because routine tasks are where a merged fact or an invented figure is easiest to miss. The checklist earns its value by being applied every time, not by having been learned once.
This course taught four small habits: decide whether an AI tool should draft the first attempt, write a clear instruction if so, check the result for what it actually says rather than just how it reads, and give it one more human check before it leaves your own team. None of the four habits is complicated on its own. Doing all four, in order, on every task, including the small routine ones, is what turns them into a reliable daily habit instead of something you only remember to do when a task feels important.
After Omar's hybrid buying update goes smoothly, a colleague suggests that weekly updates to Delta Garments are routine enough to skip the decide and check steps from now on and simply send whatever an AI tool drafts on the first try.
Should Cedar & Co. skip the decide and check steps for routine weekly updates? Decide before opening the answer.
No. "Routine" describes how often a task repeats, not whether it is safe to skip a step that catches real errors.
1. The check step caught a real error on Omar's own recent task: a sentence that merged the confirmed base quantity and the forecast top-up into one figure. Nothing about a weekly update makes that kind of merge less likely; if anything, routine familiarity makes it easier to miss. 2. The decide step is quick on a routine task, not wasted. Confirming the facts are settled and the check will be fast takes seconds once the pattern is familiar; it is not a slow step to eliminate. 3. A weekly update still reaches Delta Garments, outside Cedar & Co. The hand-off cost of a wrong figure does not shrink because the task repeats often. 4. The habit that protects the order is applying all four steps every time, not judging in advance which tasks are safe to shortcut.
The key point is that repetition is a reason to get faster at each step, not a reason to remove any of them.
Order A's own fixed commercial value remains the course's clearest reference figure for what a wrong number can be worth:
In Cedar & Co.'s hybrid buying update, a merged confirmed/forecast sentence would have told Delta Garments a single quantity that did not match either the real confirmed base order or the real forecast top-up. Delta Garments could reasonably plan production capacity against that wrong single figure, exactly the kind of costly misunderstanding this course's earlier lessons showed for Order A's own fixed USD 48,000 value. The four-step checklist costs a few minutes per task; the cost of an uncaught merged figure is measured in a factory's misallocated capacity and the time needed to correct it.
Use the T18-C02 Day-to-day AI-use checklist workbook. For one AI-assisted task this week, complete every column: the task, the facts supplied, the format requested, whether the output was fact-checked, and whether a human check happened before hand-off. Add Omar's hybrid buying update as a worked example, then repeat the full checklist on your own task from start to finish.
[Download the Day-to-day AI-use checklist workbook](../../templates/T18-C02-day-to-day-ai-use-checklist.xlsx)
1. In Omar's hybrid buying update, which step of the checklist caught the sentence that merged the confirmed base quantity and the forecast top-up into one figure?
2. A socks mill uses the same four-step checklist for a routine weekly production count sent to a domestic retailer. The count has been correct for the last ten weeks. What does the checklist recommend for week eleven?
| EN | AR | FR | TR | Meaning |
|---|---|---|---|---|
| AI-use checklist | قائمة فحص استخدام الذكاء الاصطناعي | Liste de contrôle d'usage de l'IA | Yapay zeka kullanım kontrol listesi | The ordered set of checks a person applies to any AI-assisted task, from instruction to hand-off. |
| Confirmed order portion | الجزء المؤكد من الطلبية | Portion confirmée de la commande | Siparişin onaylanmış kısmı | The part of a hybrid order that is already agreed and fixed. |
| Forecast order portion | الجزء التقديري من الطلبية | Portion prévisionnelle de la commande | Siparişin tahmini kısmı | The part of a hybrid order that is planned but not yet fixed, subject to change. |
| Instruction | تعليمة | Instruction | Talimat | The written request you give an AI tool, stating the task, the facts and the required format. |
| Human check | مراجعة بشرية | Vérification humaine | İnsan kontrolü | A person's own review of an AI-assisted output against known facts before it is used or shared. |
| Hand-off point | نقطة التسليم | Point de transmission | Devir noktası | The moment an AI-assisted output leaves the control of the person or team that produced it. |
You have completed all five lessons of Course 2. You can now decide when to ask an AI tool for a first draft, write a clear instruction, check a draft for substance, and keep a human check on any output leaving your team. Keep using the workbook checklist on every task from now on. Back to the course map