Weekly review template, run by AI
Every weekly review template asks roughly the same questions: what's done, what's stuck, what's coming up, what did I say I'd do that I haven't. The template was never the hard part. Answering those questions honestly, using information that's actually current, is. Here's the template, and what changes when the answers are already true by the time you sit down.
David Allen called the weekly review the critical success factor in Getting Things Done, and the reasoning generalizes past GTD specifically: any PKM method eventually needs a checkpoint where you reconcile what your system says against what's actually true. Below is a template that works whether you're running GTD, PARA, or nothing formal at all.
The template
| Section | The question |
|---|---|
| Clear | What's sitting in an inbox, unfiled, unread, or unanswered right now? |
| Closed | What finished this week that my system doesn't know is finished yet? |
| Stuck | What's been open for more than two weeks with no movement, and why? |
| Owed | What did I promise someone, and what's someone waiting on me for? |
| Ahead | What's coming in the next week that needs prep now, not later? |
That template is genuinely simple. The reason weekly reviews are the first habit to slip isn't the questions, it's that answering them honestly requires reconstructing the state of everything from memory and scattered notes, which takes real time and gets skipped the first busy week. That reconstruction step, not the review itself, is the part worth automating. The raw material for that reconstruction is exactly what a well-kept automated daily notes system already has sitting in it.
What changes when the pre-work is done for you
If your system already knows what closed this week because it watched the conversations where things closed, "Closed" isn't a question you answer from memory, it's a list that's already correct. If it's been tracking replies, "Owed" is already populated instead of something you have to reconstruct by scrolling through email. The template doesn't change. What changes is that by the time you sit down, every section already has a draft answer, and your job shifts from research to judgment: is this actually stuck, or just slow, is this actually done, or does it need one more step.
Template vs. workflow. This page is the questions and the structure, the thing you'd print out or paste into a note. If you want the full mechanics of how an assistant assembles the answers automatically before you sit down, that's the AI weekly review workflow, its closer sibling.
Where this pairs with other methods
The template above generalizes across methods on purpose. If you're tracking next actions GTD-style, "Stuck" maps directly onto commitments sitting in a project list with no next step. If you're running PARA, "Closed" is the moment a Project should move to Archives, covered in GTD with AI and the broader PKM methods comparison, which is why the two work best set up together rather than separately.
What a twenty-minute version looks like
Run it as a straight read-through rather than a rebuild. Open the five sections above with their answers already drafted. Skim Clear for anything genuinely urgent that slipped in today. Scan Closed and correct anything the system marked done that isn't, or missed that is. Look at Stuck and make the actual call on each one: push it, drop it, or reassign it, since deciding what to do about a stuck item is judgment no assistant should make for you. Confirm Owed matches your own sense of what you promised. Glance at Ahead and flag anything that needs prep tonight rather than tomorrow morning. That's the whole review, and it stays twenty minutes because you're confirming a draft, not writing one from scratch.
Who this is actually for
A weekly review only earns its place if you're the kind of person juggling enough moving pieces that memory alone stops being reliable, several projects, a role that spans meetings and follow-ups, a life with more open threads than a single list can hold. If that's not you yet, a lighter daily habit will do. If it is you, the review is worth the twenty minutes, and it's worth those twenty minutes being about judgment rather than reconstruction, which is the whole case covered on Claude Code for non-developers.
Sit down to a review that is already drafted
The book covers building a hub in plain files that fills in Clear, Closed, Stuck, Owed, and Ahead before you open it, so the twenty minutes go to judgment instead of reconstruction.
Get the bookClear, Closed, Stuck, Owed and Ahead, already answered
The same pre-work, built for a company: the five sections filled in from the week's actual mail and meetings, so the review is twenty minutes of judgment instead of reconstruction. Try the demo.
See the demo