Lesson 1 of 8 · 18 min
Diagnose: where AI creates value
Map your tasks, score them and choose your first three projects.
Most entrepreneurs start with AI from the tool: they open an assistant and wonder what to ask it. Results are uneven, and the enthusiasm fades quickly. This course takes the opposite path: we start from your business, not from the tool. Before you write a single request, you will know exactly where AI can save you time, how much, and where it has no place.
By the end of this lesson, you will be able to:
- take an inventory of the tasks that fill your week;
- score each task with a simple, objective grid;
- choose your first three projects and estimate the time to be saved.
The common thread of this course
Throughout the eight lessons, we will follow one case: Atelier Boréal, a ceramics studio in Montreal (a fictional business, created for this course). Its founder, Julie, works with one part-time employee. The studio sells online, runs beginner workshops on Saturdays and takes custom tableware orders from restaurants.
Julie has the same problem as many entrepreneurs: she spends her evenings on emails, posts and quotes instead of creating. Each lesson will apply the method to her case, then you will apply it to yours.
Three ways to work with AI
Before scoring anything, you need one clear distinction. For each task, AI can play one of three roles:
- Produce the first draft. The assistant writes, you review and adjust. Examples: a follow-up email, a product description, a posting plan. This is where most quick wins are.
- Support your thinking. The assistant asks questions, compares options, spots gaps, but the decision remains entirely yours. Examples: preparing a negotiation, analyzing a sales table.
- Stay out of it. Some tasks remain human: a hiring decision, a conflict with a customer, anything involving sensitive data that must not leave the business.
Step 1: the weekly inventory
During a normal week, write down every recurring task that is not your core craft. For each one: what it is, how often it comes up, how long it takes.
Don’t aim for perfection: a list of fifteen to twenty tasks is enough. If you are short on time, do it from memory using your calendar and your inbox from the last two weeks.
An assistant can help you not to forget anything:
I run a two-person ceramics studio: online sales, beginner workshops on Saturdays, restaurant orders. Help me list the administrative, sales and communication tasks that come up every week in this kind of business. Present them in a table, grouped by category, without estimating durations: I will add them myself.
Notice the end of the request: you are the one who knows the durations. The assistant provides a starting list; you correct it with your reality.
Step 2: the FTSR grid
Then score each task from 1 to 3 on four criteria:
- F · Frequency. 1 = less than once a week, 2 = every week, 3 = every day.
- T · Time. 1 = under 15 minutes, 2 = 15 minutes to an hour, 3 = over an hour each time.
- S · Structure. Does the task always follow the same pattern? 1 = different every time, 2 = partly, 3 = almost always the same form.
- R · Risk. What does a mistake cost? 1 = little and easy to fix, 2 = annoying, 3 = serious (money, reputation, personal data, legal issues).
Then calculate: Priority = F + T + S − R.
The score ranges from 0 to 8. Above 5, the task is an excellent candidate. Between 3 and 5, it deserves a supervised trial. Below 3, leave it aside for now.
Why subtract risk? Because a frequent, lengthy task where a mistake is costly requires such tight checking that the gain melts away. We will handle it later, with method.
The Atelier Boréal case
Here is part of Julie’s inventory, scored with the grid:
| Task | F | T | S | R | Priority |
|---|---|---|---|---|---|
| Answering questions about workshops (dates, prices, level) | 3 | 2 | 3 | 1 | 7 |
| Writing descriptions of new pieces | 2 | 2 | 3 | 1 | 6 |
| Preparing Instagram posts | 2 | 3 | 2 | 1 | 6 |
| Preparing quotes for restaurants | 1 | 3 | 2 | 2 | 4 |
| Following up on unpaid invoices | 1 | 1 | 3 | 2 | 3 |
| Analyzing monthly sales | 1 | 3 | 1 | 2 | 3 |
| Handling an unhappy customer | 1 | 2 | 1 | 3 | 1 |
Julie’s first three projects are clear: answers to workshop questions, product descriptions and posts. The unhappy customer stays in her hands: low structure, high risk.
Step 3: estimate the gain, carefully
For each chosen project, estimate the current time per month, then the time saved. Be careful: count on a 30 to 50% gain, no more. Reviewing, adjusting and learning take time, especially at the start.
For Julie:
- workshop answers: about 5 hours a month, estimated gain 40%, so 2 hours;
- product descriptions: about 4 hours a month, estimated gain 50%, so 2 hours;
- posts: about 8 hours a month, estimated gain 40%, so a little over 3 hours.
That is about seven hours a month on just three tasks. It is only an estimate: lesson 8 will show you how to measure the real gain and adjust course.
Common mistakes
- Starting with the most impressive task rather than the most frequent one. Saving 20 minutes every day is worth far more than saving two hours once a quarter.
- Underestimating risk. A wrong quote or an inaccurate answer to a customer costs more than the time saved.
- Sharing sensitive data without thinking: customer names, contact details, confidential figures. Lesson 8 will set a precise framework; until then, always anonymize.
- Launching everything at once. Three projects, not ten. You will add the next ones when the first are running.
Key takeaways
- Start from the business, not the tool: the inventory comes before the first request.
- AI produces a first draft, supports your thinking, or stays out of it.
- The FTSR grid (Frequency, Time, Structure, Risk) objectively ranks your tasks.
- Estimate gains carefully, 30 to 50%, then measure them.