# First AI Use Cases Starter Worksheet

<!--
HOW TO USE THIS FILE
1. Rename it to something you will recognize (for example: my-ai-use-cases-2026.md).
2. Set a timer for fifteen minutes.
3. Clear the worked example in the table below (the "Summit Mechanical" row).
4. Write down every task you repeat in a normal week, one per row.
5. Fill in the four rating columns, 1 to 5, for each task. Guidance is under the table.
6. Rank your list, pick the top three, and write a one-sentence "good result" for each.
Keep yourself in the loop the first week. Read each output against your good-result sentence and approve it before it goes anywhere.
-->

## Step 1 and 2: List and rate your recurring tasks

<!--
Rate each task 1 to 5 in the four columns:
- Repeatable: how repetitive and predictable it is (5 = you do it constantly, same shape every time)
- Structured: how structured the inputs are (5 = forms, emails, numbers, standard questions)
- Low-stakes: how low-stakes a wrong answer is and how easy it is to check (5 = easy to catch, nothing breaks)
- Minutes/wk: your estimate of minutes per week the task costs you (write the actual number, not a 1-to-5 score)
High Repeatable + high Structured + high Low-stakes + high Minutes goes to the top of your list.
-->

| Task | Repeatable (1-5) | Structured (1-5) | Low-stakes (1-5) | Minutes/wk |
|---|---|---|---|---|
| Draft first-pass replies to common customer questions (WORKED EXAMPLE, delete this row) | 5 | 4 | 4 | 120 |
| <!-- your task 1 --> |  |  |  |  |
| <!-- your task 2 --> |  |  |  |  |
| <!-- your task 3 --> |  |  |  |  |
| <!-- your task 4 --> |  |  |  |  |
| <!-- your task 5 --> |  |  |  |  |
| <!-- add rows as needed --> |  |  |  |  |

<!--
The tasks that fit best early on are predictable and repeated, with structured inputs and outputs (numbers, dates, names, statuses, standard replies). Predictability is what lets you tell quickly whether the result is right. Marketing and customer engagement is the most common first pick among small firms, but choose from your own repeated work, not a feature list.
-->

## Step 3: Define what a good result looks like

<!--
Before you try anything, write one sentence describing a result you would actually use.
Anthropic's guidance is that a good criterion is Specific, Measurable, Achievable, and Relevant.
Aim for "draft a reply a customer would actually send," not "help with email."
Write the sentence for your top pick first, then the other two once you have ranked them.
-->

**Worked example (delete before you start):**
Task: Draft first-pass replies to common customer questions.
Good result: A reply I would actually send to a customer after one quick read, in our normal tone, that answers their question and needs only a small edit.

**Your top three, ranked:**

1. Task: <!-- paste the task -->
   Good result: <!-- your one Specific, Measurable, Achievable, Relevant sentence -->

2. Task: <!-- paste the task -->
   Good result: <!-- your sentence -->

3. Task: <!-- paste the task -->
   Good result: <!-- your sentence -->

## Step 4 and 5: Try three this week, keep yourself in the loop

<!--
Treat the first week as a test, not a handoff. Three small tests beat one big automation project that stalls before it begins. Read each output against your good-result sentence and approve it before it goes out. Let a task run on its own only once it has earned your trust over a few good weeks.
-->

- [ ] Task 1 tried this week, output checked against my good-result sentence
- [ ] Task 2 tried this week, output checked against my good-result sentence
- [ ] Task 3 tried this week, output checked against my good-result sentence

<!--
Want a read on where your own skills sit today? The 7 Levels of AI Proficiency assessment takes about ten minutes: https://assess.launchready.ai
-->