The approval log
For one week, write down every piece of work that came back to you before it moved forward.
- what came back
- who sent it
- what decision they needed
- what you knew that they did not
How we think
Every lean business has repeatable work.
Leads to qualify.
Proposals to shape.
Client updates to write.
Candidates to review.
Cases to triage.
Content to approve.
Operations to coordinate.
But the work only moves properly when someone knows what matters, what to ignore, what good looks like, when to escalate, when to bend the rule, and when the context changes the answer.
In a lean business, that someone is usually the founder.
That keeps the business dependent on you even when the task itself looks delegatable.
0/7 answered
Most founders first look for tasks to automate.
But the task is usually the visible layer.
The deeper layer is the judgment around the task.
A proposal is not hard because writing is hard. It is hard because someone has to know what the client actually cares about.
A lead is not hard to score because fields are missing. It is hard because someone has to know which signals matter.
A client update is not hard because the words are hard. It is hard because someone has to know what needs confidence, what needs context, and what needs care.
If work keeps coming back to you, the business may not have an execution problem.
It may have a judgment transfer problem.
These are not homework. They are ways to see where your judgment is hiding.
For one week, write down every piece of work that came back to you before it moved forward.
Pick one piece of work you corrected.
Pick one recurring process. Write the normal rule. Then write the moments where you would break that rule.
Ask it to help you see where your thinking diverges from standard reasoning — before you try to solve the process.
Don't ask AI to solve the process yet. Ask it to help you see where your thinking diverges from standard reasoning. That gap is the judgment layer.
Based on all our conversations about my business — working through decisions, reviewing work, thinking through client situations, correcting output — I want to find every pattern where my thinking consistently diverges from yours. 1. Where have you suggested something and I corrected your direction? What was your reasoning, and what was mine? What pattern does that reveal? 2. Where have you framed something reasonably, but I rejected the framing — not just the answer, but how you were thinking about the problem? 3. Are there recurring themes in my corrections? Do I push back when you're too generic, too optimistic, too cautious, too focused on the wrong variable? 4. What do you now know not to suggest, because my corrections have shown I think about it differently? Be specific. Concrete examples from our conversations, not vague summaries. I'm trying to map exactly where standard reasoning falls short of my actual judgment — that gap is what I need to understand.
Based on everything you know about my business — from our conversations, the work we've done together, the corrections I've made — map every recurring pattern in what I flag as wrong. Not generic quality issues. The specific things that bother ME. 1. When I review work and say something isn't right, what is it usually about? Is it the wrong framing? Missing context? Too generic? Technically correct but missing the point? The wrong priority? Too cautious when the situation called for decisiveness? 2. What's the gap I most often describe between technically correct and actually correct? Where does work look done to a reasonable person, but still needs my correction? 3. Are there specific moments in a process where I consistently think people drop the ball — the opening, the diagnosis, the recommendation, the close, the follow-up? 4. What are the things I would catch that a competent person following standard process would miss? Be specific. Pull from actual corrections and feedback I've given. The patterns across them are what I need — not a single example, but the recurring shape of what I reject.
Help me map how my judgment actually flows — not what my rules are, but how they interact when I'm making a real decision. Based on everything you know about my business: 1. When I encounter a new situation, what do I assess first, and what comes later? Is there a natural sequence in how I process things? 2. Which of my decision points are gates — meaning if something fails there, the rest becomes less relevant or the answer changes entirely? 3. When two of my rules conflict — say feasibility says "this works" but experience says "this won't land well" — which one do I tend to prioritise? What does that reveal about what I actually weigh most heavily? 4. Do I ever combine multiple assessments into a single gut call rather than running them separately? Which ones merge, and what's the combined judgment? 5. Does my sequence change based on the type of client, project, or situation? How? Think about this from actual decisions you've watched me make. Not theoretical order — the real flow as you've observed it.
Make sure memory is on in your ChatGPT or Claude settings. Don't use ChatGPT or Claude often? That's fine — let's talk, and I'll walk you through how to do this.
Add rough numbers. The goal is not precision — it is to see whether the bottleneck is expensive enough to map.
$5,000
$2,500
$30,000
We start with the bottleneck.
Where does work pause?
What keeps coming back?
What looks done but still needs you?
What changes your answer from yes to not yet?
Then we look for the smallest useful unit.
One recurring loop. One judgment-heavy process. One place where work could move with less founder intervention.
Judgment is hard to describe in the abstract.
It is easier to see in corrections, approvals, exceptions, and edge cases.
That is where we start.
You do not need to know exactly what to build before the call.
You only need to know where the business slows down because your judgment is still required.