# Where could AI help this week?

Score up to 15 recurring tasks by minutes a week, how repeatable each is and how easily you can check the result. The three highest scores are the tasks to try with AI first.

Tools · v1.1 · draft · https://hayzlabs.com/tools/where-could-ai-help-this-week/

In short:
- List your recurring tasks and score each as minutes a week times repeatability times checkability.
- Try the three highest scores first, one at a time. Sensitive tasks are listed apart, and a task you cannot check scores 0.
- The score only says where to try first, so time each try with and without AI before you trust a saving.

Score the tasks you repeat in a week by how long they take, how alike each run is and how easily you can check the result. The top three are the ones to try with AI first. It takes about 10 minutes, in the hands-on tool below or on paper.

## What to enter

<!-- WIDGET whereAiCouldHelp -->

One row per recurring task, up to 15.

| Input | Unit | Hint |
|---|---|---|
| Task | words | A verb and an object, such as "write the weekly status report". |
| Minutes a week | minutes | Your best guess, or last week's calendar. |
| Repeatability | 1, 0.5 or 0 | 1 if the steps are the same each time, 0.5 if they are similar, 0 if each run is different. |
| Checkability | 1, 0.5 or 0 | 1 if you can check the result in a minute, 0.5 if checking takes care, 0 if it is hard to check. |
| Sensitive | yes or no | Yes if the task involves other people's private data or anything you should not paste into a chat. |
| Hoped saving | minutes a week, optional | What you hope AI will save on this task. |
| Hours in your week | hours | 40 unless you work more or fewer. |

## Reading the answer

| Result | What it means |
|---|---|
| Top three | The three highest scores above zero. Try them one at a time, in that order. |
| A score of 0 | The task is different each time, or you cannot check it, or both. It is a poor first try. |
| Listed apart | A sensitive task. It is not ranked. Read [What not to paste](https://hayzlabs.com/start/what-not-to-paste/) before you put any of it in a chat. |
| Optimistic | Your hoped savings add up to more than a quarter of your week. Lower them and run it again. |
| Fewer than three | Fewer than three tasks scored above zero. Add more tasks, or accept that few of yours fit. |

> [!OUR RULE]
> The score is minutes a week times repeatability times checkability. If your hoped savings add up to more than 25% of your week, the page says "optimistic". The score, the three levels and the 25% line are ours, and no study tests them.

Multiplying means a task you cannot check scores 0 however long it takes. That is deliberate.

For context on the 25% line, US workers who use generative AI (tools that write text or make images) report saving about 5% of their hours, roughly two hours in a 40-hour week (self-reported, Nov 2024). Across all workers the saving is about 1.4% ([St. Louis Fed](https://www.stlouisfed.org/on-the-economy/2025/feb/impact-generative-ai-work-productivity), survey).

## Worked example (illustrative)

Six invented tasks for one week of 40 hours.

| Task | Minutes a week | Repeatability | Checkability | Score |
|---|---|---|---|---|
| Weekly status report | 60 | 1 | 1 | 60 |
| Meeting notes | 90 | 1 | 0.5 | 45 |
| Inbox triage | 150 | 0.5 | 0.5 | 37.5 |
| Expense filing | 30 | 1 | 1 | 30 |
| Client proposals | 180 | 0.5 | 0 | 0 |
| Contract summaries (sensitive) | 45 | 1 | 0.5 | Listed apart (it would score 22.5) |

The top three are the weekly status report, the meeting notes and the inbox triage. The proposals take the most time of any task but score 0, because the result is hard to check.

## What to do with the answer

Try the first task with and without AI and time both runs, including checking, before you believe any saving: [Am I actually faster?](https://hayzlabs.com/tools/am-i-actually-faster/).

> [!LIMIT]
> - The score ranks where to try first. Only your own runs show whether AI works on the task.
> - Your minutes are guesses. A week of notes beats memory.
> - A low score does not mean AI cannot help. A task that differs each time can still be worth a draft or a second opinion.

Nothing you record on this page is stored or sent anywhere, and the tool keeps nothing between visits.

## Sources

1. St. Louis Fed. The Impact of Generative AI on Work Productivity. Survey, self-reported, Nov 2024. [Link](https://www.stlouisfed.org/on-the-economy/2025/feb/impact-generative-ai-work-productivity)