Run your day · Lesson 6.6
Monday Automation vs Doing It by Hand: Which Is Worth It?
You do the same job every Monday. You could keep doing it by hand, or you could spend an afternoon setting up Monday automation so AI does it for you. Which one is worth your time?
The honest answer is that it depends on the job, and the numbers you need are your own. This post gives you a way to compare the two routes fairly, without invented figures, so you can decide before you invest an afternoon.
Lesson 6.6 walks you through it step by step: Hand off the job you do every Monday.
Automation pays when the job is frequent, stable and cheap to check; the manual way wins when it changes every week or when a mistake is costly.
What you are really comparing
People compare "an hour by hand" with "zero minutes automated". That is not the real comparison. The real one has four parts.
- Setup: writing the steps, testing, checking the first runs.
- Running: the time you still spend reading notes and approving output.
- Maintenance: fixing it when a file moves, a column changes or a tool updates.
- Risk: what happens if it is wrong and nobody notices.
Manual work has almost no setup and no maintenance, but all of its cost is in the running. Automation flips that: the cost is up front and in upkeep.
The comparison table
| By hand | Automated with AI | |
|---|---|---|
| Setup | None | Real: write steps, test, check three runs |
| Weekly effort | Full job each time | Reading a short completion note |
| Consistency | Varies with your mood and rush | Follows the written steps, though it can drift |
| Handles odd cases | Yes, you adapt | Only if the steps cover them |
| Failure mode | You skip it or rush it | Fails quietly if you stop checking |
| Best for | Changing, high stakes work | Fixed, frequent, low stakes work |
Run your own numbers
You can work out the answer with a pen and a few honest estimates. Do not guess what AI saves in general. Measure your job.
- Time the manual jobNext Monday, time it from start to finish. Use the real number, not the number you remember.
- Estimate setupBe generous. Writing steps, one test and three checked runs will take more than you expect.
- Estimate weekly upkeepReading a note and fixing the odd problem. Add a little each month for breakage.
- Compare over a quarterThirteen weeks of the manual job against setup plus thirteen weeks of upkeep.
- Add the risk questionIf it went wrong once and you found out three days later, how bad would that be?
If the quarter comparison is close, do not automate yet. The point of the exercise is to find jobs where the gap is clear.
A worked example, illustrative only
These numbers are made up to show the method. Say you run a small bakery and every Monday you copy weekend orders from a sheet into a prep list. Suppose it takes forty minutes by hand. Suppose setup takes three hours and weekly upkeep is five minutes.
Over thirteen weeks, the manual route costs thirteen times forty minutes. The automated route costs the three hours of setup plus thirteen times five minutes. In this invented case automation wins well before the quarter ends. Change the numbers, such as a job that takes ten minutes, and the answer flips. That is why you should use your own.
Where the manual way still wins
- Jobs that change shape every week, where the steps cannot be written down.
- Work that needs your judgment about a relationship, such as a delicate client email.
- Anything where an error cannot be undone, such as a payment or a deletion.
- Jobs you do rarely. Setup never pays back.
There is no shame in keeping these by hand. A good rule is to automate the preparation and keep the decision. The AI drafts, you decide.
Middle routes
You do not have to choose all or nothing. A repeating calendar event, which Google Calendar supports, gives you a prompt without any automation at all. Or you can automate half the job: have AI prepare the list and do the final step yourself. The Hermes Agent documentation, for example, describes scheduled tasks that deliver their results back to you, which fits the "prepare and I will review" pattern.
Our checklist for choosing among scheduling options is in the tools post.
A decision checklist
If you want a quick answer before doing the full sums, run through this list. Count the yes answers.
- Do I do this job at least once a week, most weeks of the year?
- Are the steps the same each time, even when the content changes?
- Could I write the steps in ten lines or fewer?
- Would I notice a wrong result within a minute of reading it?
- Would one wrong result be an annoyance, not a crisis?
- Do I dislike the job enough to keep going through setup?
Five or six yeses: this is a strong candidate, so do the quarter comparison and probably go ahead. Three or four: automate part of it, such as the preparation, and keep the final step by hand. Two or fewer: leave it manual for now and revisit when the job settles down.
The checklist is not a formula. It is a prompt to think about the parts of the decision people skip, especially stability and risk. Most regrets about automation come from picking a job that failed questions two and five, not from the technology.
Also remember the cost of attention. Even a job that runs on its own asks for a little of your attention each week to read the note. If your time is already stretched thin, that small tax still counts.
Decide, then check
Pick a job that passes the quarter test and the risk test. Run it by hand with AI first, then watch three runs. If the first month goes badly, go back to doing it manually and you have lost an afternoon, not a week. Lesson 6.6 walks you through the whole path: Hand off the job you do every Monday.
Questions people ask
Is automation always faster?
Not in the first weeks. Setup and checking take real time, and the gain comes later if the job is stable.
What if the AI makes a mistake?
Plan for it. Use draft only rules, read the completion note, and keep the manual method in reach.
Should I automate a job I do once a month?
Probably not. Rare jobs rarely repay the setup.