The Monday Morning Audit
Five questions to ask yourself before you open a single AI tool — and what the answers tell you.
Last week we talked about the Sunday reckoning — that moment when your brain runs the math on a week of producing without owning. If you read that and felt something land, this week is the follow-up.
Because here's what happens after a Sunday reckoning: Monday morning comes, and you have a choice. You can start the week the way you ended the last one — with the AI tools firing before you've even located yourself — or you can take five minutes first.
This is not a productivity hack. This is a diagnostic.
Five questions. No one's watching. Answer honestly.
The Monday Morning Audit
Did you make a decision yesterday that wasn't AI-suggested?
Not a preference — a real decision. Something where the answer could have gone another way, and you had to actually think through it.
What it reveals: If the answer is "no," that's not a character flaw. It's a signal that your decision-making muscle hasn't been used in a while. That's recoverable — but only if you notice it.
When you think about the code you shipped last week — what's the texture?
Pride? Numbness? "It works"? Something else?
What it reveals: Engineers who've been through the AI fatigue cycle consistently describe the same thing: the work they shipped started to feel like it was happening to them rather than through them. If "it works" is as far as the feeling goes — that's worth naming.
What's your cognitive baseline today — compared to three months ago?
Not your best day versus your worst. Your average Tuesday.
What it reveals: Most engineers don't track this consciously — but they feel it. The comparison that's actually useful isn't "am I tired today?" It's "was I this tired on a random Tuesday in January?" If the baseline has dropped, that's data, not weakness.
What's the thing you're about to do today that you already know how to do?
Not "what's on your plate." What's the task where you've already solved this before, or know exactly how to solve it — and AI would make it faster?
What it reveals: The tasks we already know how to do are the most seductive to hand off to AI — and the most costly to lose. Speed on the known is a trade. Make it deliberately, not by default.
What would you say if someone asked you how you're actually doing?
Not "fine." Not "busy." What would you actually say?
What it reveals: This one's not about AI directly. It's about whether you still have access to an honest self-assessment. If you can't answer this without sliding into deflection, that's worth knowing.
What to Do With the Answers
You don't have to fix anything today. That's not the point of an audit — the point is to have accurate information.
But here's what tends to happen when engineers sit with these five questions honestly: they realize they're managing something they've been too busy to name. And naming it is the first move, not the last.
If Question 2 landed hard — if the texture of your recent work is "numb" or "it works" — that's a prompt to look at the recovery guide. Not because something is broken, but because your relationship with your own work is worth paying attention to.
If Question 4 pointed at something you actually want to keep knowing — a skill that matters to you — that's a prompt to try a no-AI session this week. Not as a detox cleanse. Just as a recalibration.
A Framework Worth Knowing
Most teams have never explicitly evaluated which AI workflows preserve capability and which ones erode it. The AI Decision Stack is a simple 4-layer framework for making that evaluation team-wide — cognitive cost, skill preservation, team dynamics, long-term sustainability.
It's not a policy. It's a vocabulary. And most engineers who find it say they wish they'd had it months ago.
The Three-Question Close
If you don't have time for a full audit, try these instead:
1. Did I make anything mine today?
2. Am I thinking clearly right now, or just productive?
3. Do I want to do this work, or just finish it?
Three "just finish it" or "neither" answers in a row is worth paying attention to. Not as a crisis — as a data point.
If this landed for you, here's where to go next:
Take the AI Fatigue Quiz Recovery Guide AI Decision StackYou don't have to have answers today. You just have to know where you're standing.
— The Clearing