The Weekly Productivity System

A week-long rhythm with AI in the loop. Four touchpoints, none longer than fifteen minutes.

This is deliberately small. Productivity systems fail because they are too elaborate to sustain, not because they lack features.

The four touchpoints

When	What	Time
Monday morning	Plan the week	15 min
Each morning	Brief the day	5 min
Each evening	Close the day	3 min
Friday afternoon	Review the week	15 min

Monday: plan the week

Dump everything, then let AI sort it.

Here is everything on my plate this week:
[BRAIN DUMP - meetings, deadlines, half-finished things, nagging tasks]

My fixed commitments: [MEETINGS AND THEIR TIMES]
Focused hours realistically available: [N]

Sort into:
1. Must happen this week, with the real consequence if it does not
2. Should happen
3. Can wait — and say until when

Then flag:
- Anything urgent but not important
- Anything I have carried for more than two weeks (why is it still here?)
- Anything that will take longer than I think

Finally: propose an order given my [N] focused hours. Be realistic —
if it does not fit, tell me what to drop rather than compressing it.

The "tell me what to drop" instruction is what makes this useful. A plan that assumes you have more time than you do is just a list.

Each morning: brief the day

Today is [DAY]. My plan for the week was: [PASTE MONDAY OUTPUT]
Yesterday I finished: [WHAT]
Yesterday I did not finish: [WHAT]
Today's meetings: [LIST]

Give me:
- The three things that must happen today
- The one thing that matters most, and when to do it given my meetings
- Anything from yesterday that has now become urgent
- One thing I could drop if the day goes badly

Under 120 words.

Each evening: close the day

Three lines, typed straight into a note:

Done:
Not done, and why:
Tomorrow's first task:

That is the whole thing. The "why" column is the valuable part — after two weeks you will see the same reason repeatedly, and that reason is your actual productivity problem.

Friday: review the week

Here are my daily closes for this week: [PASTE]
Here was Monday's plan: [PASTE]

Tell me:
1. What did I plan and not do? Is there a pattern in what slips?
2. What took much longer than expected?
3. What did I do that was not on the plan? Where did that come from?
4. Which of my "must happen" items actually mattered by Friday?
5. One change for next week. Just one.

Be direct. I am not looking for encouragement.

Point 4 is uncomfortable and worth it. Most weeks, at least one urgent thing turns out not to have mattered.

The AI-assistable audit

Once a month, work out what to automate next.

Here are my daily closes for the last month: [PASTE]

Which recurring tasks appear most often?
For each, classify:
- AI could do most of it, I review
- AI could draft it, I rewrite substantially
- AI cannot meaningfully help

For the first category, rank by total time spent per month.

The top of that list is your next automation, and you now have evidence rather than a hunch.

Making it stick

• Put the prompts somewhere reachable in two seconds. A note, a text expander, a pinned document. If it takes thirty seconds to find the prompt, you will stop.
• Do the evening close even on bad days. Especially on bad days — that is where the pattern is.
• Skip the Friday review twice and the system is dead. It is the only part that changes anything.

What this deliberately does not do

• It does not track every task. Use whatever task list you already have.
• It does not measure your time in detail. The daily close is enough signal.
• It does not gamify anything.

The whole system is four prompts and a three-line note. That is the point.

Your setup

Where my prompts live
Where the daily close goes
Monday planning time
Friday review time
One change I am trying this week