Why Decisions Still Come Back to You After You Delegate
Why Decisions Still Come Back to You After You Delegate
You hired help. You wrote the process. You handed off the actual doing. And the same handful of decisions still land on your desk every week, like what to do with a difficult customer or what goes in an unusual quote. That is not a delegation failure. You delegated the task and kept the decision, and those are two different things.
Why doesn't handing off the task stop the questions?
A task is a set of steps. A decision is the judgment call inside those steps. Handing someone the steps, like send the invoice or answer the message, does not tell them what to decide when the situation isn't the standard one. So they do the obvious parts and bring you the parts that aren't, every time, because nobody ever wrote down what you actually decide in those moments. You never handed off the deciding. You handed off the parts around it.
What decisions are actually stuck on you right now?
Most owners can't list them off the top of their head. The fastest way to find them isn't to guess, it's to ask. Sit down with two or three people who work for you and ask directly: which decisions do you feel are genuinely yours to make, and which ones do you still bring to me out of habit rather than because you actually need my answer? Write down exactly what they say, not what you assume they'd say. Run it with two or three people and a pattern shows up fast. It's usually the same one or two decision types, repeated across different people, that nobody ever actually assigned to anyone.
The ones that show up almost everywhere:
What discount or exception a difficult customer gets
How a tricky or upset message actually gets answered
What goes into a quote when the job isn't standard
Whether a late payment gets chased or let slide
How do you turn one of those into a rule instead of a question?
Name the decision plainly. Not "customer service," but "what we do when a customer asks for a discount we don't normally give."
Pull the last five or six real times it came up. Actual cases, not hypotheticals.
Write the rule as a plain if-then based on those real cases. If the order is under a set size, say yes. If it's a repeat customer, say yes. Anything bigger comes to you, and only that comes to you.
Test the rule against the next few real cases before you trust it. If it holds, hand it off with the task. If it doesn't, the rule was wrong, not the idea.
The task got handed off months ago. The decision never did.
What changes once the rule actually exists?
Once a decision is written down as a plain rule, it stops mattering who or what is running it. A new hire can follow it without calling you first. And it's also exactly the kind of thing an AI agent can run without inventing an answer, because you gave it the actual rule instead of asking it to guess your judgment. An agent trained on your own content and your own past decisions can follow that same rule the same way a person would.
Our lead magnet demo is built around exactly that idea: feeding an agent your own answers instead of asking it to guess them.
These stuck decisions usually show up first as a growing backlog, more on spotting that in how to tell if your bottleneck is sales or operations. And writing the actual rule down is the same skill as writing instructions for an agent, covered in how to write AI chatbot instructions that actually work.
Frequently Asked Questions
What's the difference between delegating a task and delegating a decision?
A task is the steps someone follows. A decision is the judgment call inside those steps, like what exception to make or how to word something touchy. Handing off the steps without writing down the judgment call means the judgment call still comes back to you every time.
How do I find out which decisions are actually stuck on me?
Ask two or three people directly which decisions they think are genuinely theirs versus which ones they still bring to you out of habit. Write down their answers, not your guess. The same one or two decision types usually show up across all of them.
Do I need software to write a decision rule, or can I just write it down?
Write it down first. A plain if-then on paper or in a shared doc is the whole fix for most decisions. Software or an agent only matters once the rule already exists and works. You're not skipping the writing-it-down step no matter what you use after.
What if a decision genuinely needs my judgment every time?
Some do. The goal isn't to remove yourself from every decision, it's to stop the ones that only feel like they need you because nobody ever wrote down the pattern. If you test a rule against real past cases and it keeps needing exceptions, that one probably does need to stay with you.
Can an AI agent actually handle a decision like this?
Only the ones you've already turned into a written rule. An agent trained on your actual content and past decisions can apply a rule you wrote consistently. It can't invent your judgment for a decision you never wrote down in the first place.
Comments