A buyer's own daily work includes drafting communication to a factory, comparing versions of a tech pack, summarising a long email thread before a meeting, and tidying notes into a clear document. Some of this work is a strong match for what an AI tool actually does well; some of it is not, and treating every task the same way wastes the tool's real strengths or creates a real risk.
This lesson gives you a working list of the task types where an AI tool's own behaviour, producing fluent, well-organised text from patterns, is a genuine advantage on an export order like Order A. Without this list, a buyer either under-uses a tool that could save real time on routine writing, or reaches for it on a task it was never suited to.
You need: the working idea of what an AI tool does from lesson 1.1 is useful preparation but not required.
By the end of this lesson you can:
Not covered here: a ranked list of AI products, or a claim about how accurate any specific tool is. Lesson 1.3 covers the mistakes an AI tool still makes on these same kinds of tasks.
| Term | Meaning |
|---|---|
| Drafting | Producing a first version of a document or message that a person will still check and finish. |
| Summarising | Reducing a longer piece of text to its key points without changing their meaning. |
| Tidying | Reorganising or clarifying existing text without adding new facts to it. |
| Comparison task | Setting two pieces of text or two versions of a document side by side to find what changed. |
| Fact-lookup task | A task whose answer depends on a specific, checkable fact rather than on organising existing text. |
The tasks an AI tool is best at share one property: the information needed to do them well is already present, either in the text you supply or in general, well-known patterns of how a document like this is normally written. Drafting a routine email, summarising a long thread, or tidying a messy paragraph all fit this description. A task that instead depends on a specific, checkable fact about your own order, one the tool was never given, is a poor match, no matter how confidently the tool answers.
| What can differ | The shapes you will meet | Why it matters to you |
|---|---|---|
| Export order under Make to Order, buyer side | Sara at Northline Apparel drafts a routine query to Delta Garments about a tech pack detail, supplying the relevant lines herself. | The task is mostly organising a clear question from information Sara already has, which is exactly what an AI tool does well. |
| Domestic Make to Stock replenishment | A domestic brand's buying team summarises several suppliers' weekly reorder confirmations into one short overview. | Summarising several already-received messages is a strong match; inventing a supplier's own confirmed number would not be. |
| Hybrid order handled through an agent | An agent drafts a comparison of two buyers' packing requirements from two documents supplied to the tool. | Comparing two supplied documents is a strong match; asking the tool to guess which requirement is correct without the documents is not. |
| A first-time task compared with a repeated one | The first time a buyer drafts a claim letter, they check every sentence closely; after several uses, checking can become rushed. | The task type does not change, but the risk of a skipped check rises as the task starts to feel routine. |
| Different product families | A socks or home-textile buyer summarising a supplier's fault-rate note faces the same task shape as Sara's tech pack query; only the product vocabulary changes. | The underlying rule, organise or restate supplied text rather than invent a fact, holds across every product family. |
Northline Apparel has confirmed Order A with Delta Garments: 15,000 T-shirts, single jersey 180 GSM, Ne 30/1, agreed FOB USD 3.20 per piece, total order value USD 48,000. Before a call with Mona, Sara wants a short summary of a 12-message email thread covering the approval history of Order A's specification, so she does not have to re-read every message.
Sara pastes the full thread into an AI tool and asks for a short summary of what was approved and when. The tool produces a clear, ordered summary using only the content of the thread she supplied. Sara checks the summary's key dates and approval points against the original messages before the call, confirms they match, and uses the summary as her own speaking notes.
The worked numberThis task suits the tool because every fact in the summary already existed in the thread Sara supplied; the tool only had to organise it, which is what it is genuinely good at, and that is exactly why the time saved here is real rather than only apparent.
The methodA common mistake is asking an AI tool a direct question about a business fact, such as a supplier's confirmed price or a specific tech pack measurement, instead of supplying the document that contains the true answer and asking the tool to summarise or extract from it. The first approach invites a confident but unverified guess; the second gives the tool exactly the kind of task it does well. The cost of the first approach is a wrong fact repeated in good faith, which then needs correcting after it has already reached someone else.
Think of an AI tool as an assistant who reads fast and writes clearly, but who was never in the meeting where your order's facts were agreed. Give that assistant the notes, the thread or the document, and ask them to organise, shorten or tidy it, and they will do a genuinely good job quickly. Ask them instead to remember a fact from a meeting they never attended, and they may still answer confidently, even though they have nothing real to base it on.
Sara is preparing for a call and, instead of emailing Mona to ask for Order A's current stage, types a question into an AI tool: "What stage is Order A at right now at Delta Garments?"
Should Sara use the AI tool's answer in the call with Mona? Decide before opening the answer.
No. Sara should send Mona a one-line message asking for the confirmed stage instead of using the AI tool's guess.
1. The AI tool was never given Delta Garments' real schedule, so its answer is a plausible pattern, not a confirmed fact about Order A. 2. This is a fact-lookup task, not an organising task, exactly the kind of question this lesson's rule says an AI tool is a poor match for. 3. A one-line message to Mona takes less time than the risk of stating a wrong stage to her own management if the guess is repeated as fact.
The AI tool behaved exactly as expected: it produced a fluent, plausible answer to a question it had no real information to answer. The mismatch was in the question Sara chose to ask it, not in the tool's own behaviour.
Order A's total order value is fixed at:
A well-matched task, such as Sara's thread summary, saves real minutes with low risk, because every fact in the result was already supplied. A poorly matched task, such as asking the tool to guess Delta Garments' current stage, risks a wrong fact reaching a meeting about an order worth USD 48,000, which can cost far more in corrected expectations and lost time than the minutes the tool ever saved.
Use the T18-C01 AI use worksheet. In the Task fit sheet, record each task you gave an AI tool, whether it was built from text you supplied or depended on a fact the tool did not have, and whether the outcome needed a correction. Start with Sara's own thread-summary example, then repeat the exercise for a task from your own week.
[Download the AI use worksheet](../../templates/T18-C01-ai-use-worksheet.xlsx)
1. Why does Sara's task of summarising the 12-message thread suit an AI tool well, while asking it Delta Garments' current production stage does not?
2. A buyer at Fairway Stores asks an AI tool to draft a comparison of two suppliers' packing instructions, supplying both suppliers' actual documents in the request. Is this task a good match for an AI tool, and why?
| EN | AR | FR | TR | Meaning |
|---|---|---|---|---|
| Drafting | صياغة مسودة | Rédaction d'un brouillon | Taslak hazırlama | Producing a first version of a document or message that a person will still check and finish. |
| Summarising | التلخيص | Résumé | Özetleme | Reducing a longer piece of text to its key points without changing their meaning. |
| Tidying | التنظيم | Mise en forme | Düzenleme | Reorganising or clarifying existing text without adding new facts to it. |
| Comparison task | مهمة مقارنة | Tâche de comparaison | Karşılaştırma görevi | Setting two pieces of text or two versions of a document side by side to find what changed. |
| Fact-lookup task | مهمة البحث عن حقيقة | Tâche de recherche de fait | Gerçek arama görevi | A task whose answer depends on a specific, checkable fact rather than on organising existing text. |
Next you will learn why the same behaviour that makes an AI tool useful for drafting and summarising also causes specific, predictable kinds of mistakes, seen through a domestic stock programme's own data. Back to the course map