Health & time · Recipe

Weekly review in 20 minutes.

Close the week just gone. Set up the week ahead. A two-mode interview and a plain-text output you'll actually paste into your notes.

Time: ~20 min Cadence: weekly (Friday or Sunday) For: anyone whose weeks blur into each other Cost: free tier of any chat AI

The recipe

1. Have your calendar open

You'll want to be able to glance at last week and next week. Notes app too, if you keep one — but not open in a way that makes you edit it.

2. Paste this opener

The opener runs two interviews back-to-back — looking back, then looking ahead — and produces a short document you can paste into whatever you actually use.

You're going to run me through a 20-minute weekly review. Half of it is looking back at this week. Half is looking at next week. HOW TO WORK WITH ME - Ask one question at a time. Short questions. - If I say "skip", move on. If I say "you decide", make a reasonable guess and move on. - Don't summarise my answers back to me. Just the next question. - When you switch from looking-back to looking-ahead, tell me plainly: "OK — now next week." WHAT YOU NEED FROM ME Looking back at the week just gone: 1. What actually got done that mattered. 2. What I said I'd do and didn't. 3. What I noticed about how I was working — energy, focus, mood, anything drifting. Looking ahead to next week: 4. Roughly what's on the calendar. I'll paste titles only, no attendee names. 5. The one or two things that, if they get done, would make it a good week. 6. Anything I can move, drop or hand off to make those things possible. WHEN YOU'VE GOT ENOUGH Produce a short plain-text document. Nothing fancy — I'm going to paste this into a notes app. In this order: Week just gone - What worked (up to 3, one line each) - What didn't (up to 2, one line each) - Note to self (one sentence) Next week - The 1-2 things that matter (verbs — what I'll actually do) - Watch out for (calendar clashes, energy risks, things I know I'll be tempted to skip) - Monday's first move (one small concrete thing to set the tone) Keep it lean. No headings other than those. No emoji. No summarising why weekly reviews are useful. Start when I'm ready.

3. Answer honestly, in either direction

The review is only useful if you tell it the drift, not the highlight reel. "I said I'd read the report and didn't" is more useful than "great progress on the strategy piece" even if both are true.

4. Paste the output somewhere you'll actually look

Notes app, journal, back of a notebook. Not the chat. The chat is a workspace, not a home for the document.

What done looks like

Six short blocks of text on one screen: three "what worked", two "what didn't", one note to self, one or two things that matter next week, a couple of things to watch and one Monday move. You'd read it back in 30 seconds.

House bar · 1 of 3

Data quality

Reviews are the format most easily corrupted by wanting to look good. The model can't tell if you're bluffing. You have to not bluff.

Say the drift out loud. "Energy was low from Wednesday." "I lost Thursday to the inbox." That's the data the review is for.
Distinguish "didn't do" from "chose not to". Both are fine. The first goes on next week's radar; the second is closed and doesn't.
If next week's "one or two things that matter" are the same as this week's "didn't do", say so out loud. That's a signal, not a coincidence.

House bar · 2 of 3

Security

Calendars leak a lot about other people. Some care needed.

Paste calendar titles, not attendees. "1:1 with report" is fine. "1:1 with Sarah about her promotion" is not.
Don't name clients, customers or patients. If it matters for the review, use a generic ("client A", "the pharma account").
Personal account, not work. Even if the review is about work, the output is about you — and the notes-to-self line is not something you want in a corporate transcript.

House bar · 3 of 3

Test it worked

A good review is one that makes next week smaller, not fuller. That's the whole check.

Look at "next week: the 1-2 things that matter". If there are three, cut one. If there are four, the review didn't do its job. Ask for a v2.
Look at "what didn't" and "next week". Did anything from "didn't" reappear on next week without a reason? If yes, you're setting up to fail again. Either fix the conditions or drop it.
Would you actually paste this into your notes app right now? If not, it's too long or too florid. Ask for tighter.

Why this works

Four things are doing the work. Weekly reviews fail more often than they succeed — usually because they're too grand. This shape keeps them small.

Two-mode interview. Same opener, two conversations. The explicit "OK — now next week" transition tells your brain to switch modes too. That switch is where most home-grown reviews trip.
Portable output. Plain text, no headings other than the fixed ones, no emoji. The best review is the one you'll actually keep — that means paste-into-notes-shaped from the start.
An anti-optimism gate. "The 1-2 things that matter" plus the "watch out for" prompt forces you to name the risk. Left to itself the review would end with a to-do list you already knew you'd blow through.
Escape hatches for a tired Friday brain. "Skip" and "you decide" mean the review gets finished. A half-finished review is worse than none — it teaches you that reviews don't complete.

Bend it by changing the fixed sections. Add a "someone I'd like to thank" line if you want the review to feed a habit of appreciation. Add "energy score 1-10" if you want a rough time series. Same interview, different notes.

Learn the underlying skill

The modules that map to this recipe: