Good list. I recommend briefly capturing the answers aka creating a reusable doc, which helped me improve at delegating over time. And adding info like "this is best used for/when..." and "alternatives include..." can help delegatees become even more independent.
Pausing to ask the right questions before handing something off has saved me from a lot of rework. Delegation feels less scary when the thinking is done up front.
At Embarc, the account teams who wrote things down even when nobody asked always shipped less rework. Half of delegation is just making the ambiguity visible before someone else has to guess at it.
I'd add a seventh: what can they decide without asking me? We once gave 1,528 call center reps the authority to resolve customer issues up to $2,000 each, with live expert backup and managers who honored every decision inside the limit. The average came to about $579 a year. Your question 2, the why, is what made that number safe.
The mise en place framing for question 3 is the one I’m stealing — that’s the step most people skip.
The line about this working for AI too is where I’d add something. Five of the six transfer cleanly. Question 2 doesn’t, quite.
“Why are we doing this?” works with a person because they carry the why forward into judgment calls you never see. A model takes the why as input and produces better output, but nothing persists. You still own every call it implies.
In coaching I see that gap surface later — leaders who delegated the task and quietly delegated the judgment with it, then can’t reconstruct why they chose what they chose.
Curious whether your checklist changes when the thing you’re delegating to can’t be accountable.
The mise en place framing for question 3 is the one I’m stealing — that’s the step most people skip.
The line about this working for AI too is where I’d add something. Five of the six transfer cleanly. Question 2 doesn’t, quite.
“Why are we doing this?” works with a person because they carry the why forward into judgment calls you never see. A model takes the why as input and produces better output, but nothing persists. You still own every call it implies.
In coaching I see that gap surface later — leaders who delegated the task and quietly delegated the judgment with it, then can’t reconstruct why they chose what they chose.
Curious whether your checklist changes when the thing you’re delegating to can’t be accountable.
“It will take too long to explain how to do this” might be one of the most expensive sentences an expert can repeatedly say.
I work mostly with deep subject matter experts, and I hear a version of this all the time.
They're right. Initially, it probably IS faster to do it themselves. Explaining “common sense” to someone else can feel frustrating.
So they do it themselves. Again!
Eventually they become the go-to person. Take a number. Get in line. Which can feel like success, but it really isn't.
I've written before about knowledge transfer as an important expert capability, and I think your question “why are we doing this?” is particularly important. Never underestimate the power of the "why".
I would suggest a different take on your questions. They are great questions. But rather than asking them of yourself before you delegate, could you instead take a coaching approach and ask them?
What do you know about this already? How would you go about it?
Why do you think this is important?
What will you need?
What do you think great might look like?
What is a reasonable timeframe to complete this?
What are you worried could go wrong? How might you overcome that?
By asking questions, you turn them from passive task-takers into proactive learners. They are more likely to truly understand the task if you ask them to think through it themselves, with your guidance as a coach.
I would add one concrete example of an acceptable output and name the person who handles exceptions. Otherwise the receiver has to assume both quality and authority.
Delegation gets easier when the handoff includes a definition of done, one example of acceptable output, and the first check-in date. That small upfront cost prevents the manager from becoming the bottleneck.
I spent 27 years on the NYSE floor interviewing seasoned CEOs on TV. I noticed some answered a different question. The question they wanted to answer, versus the one I asked. The audience usually knows.
When I coach senior women today, the failure isn't a confidence problem, it's usually a structural problem: She hasn't pre-built the three sentences.
A no without architecture around it reads as “Oh she’s being difficult” A no with architecture around it reads as graceful and still preserving the relationship.
The transfer of judgment is what I see most people underestimate. The hard question - the one that decides whether the delegation actually works - is how much context goes along with the task. I've seen this in EdTech too: the briefing that felt thorough rarely was.
Number 4 is the one I only learned through AI. I had a whole page of rules describing the tone I wanted and the drafts kept coming back dull. Then I pinned one message I'd already approved at the bottom of the rules and said match this one. Fixed it immediately, which was annoying after all those rules.
Good list. I recommend briefly capturing the answers aka creating a reusable doc, which helped me improve at delegating over time. And adding info like "this is best used for/when..." and "alternatives include..." can help delegatees become even more independent.
Pausing to ask the right questions before handing something off has saved me from a lot of rework. Delegation feels less scary when the thinking is done up front.
At Embarc, the account teams who wrote things down even when nobody asked always shipped less rework. Half of delegation is just making the ambiguity visible before someone else has to guess at it.
Super actionable post! I've turned it into a Claude Skill ;)
Thanks for pulling this together and sharing.
I'd add a seventh: what can they decide without asking me? We once gave 1,528 call center reps the authority to resolve customer issues up to $2,000 each, with live expert backup and managers who honored every decision inside the limit. The average came to about $579 a year. Your question 2, the why, is what made that number safe.
The mise en place framing for question 3 is the one I’m stealing — that’s the step most people skip.
The line about this working for AI too is where I’d add something. Five of the six transfer cleanly. Question 2 doesn’t, quite.
“Why are we doing this?” works with a person because they carry the why forward into judgment calls you never see. A model takes the why as input and produces better output, but nothing persists. You still own every call it implies.
In coaching I see that gap surface later — leaders who delegated the task and quietly delegated the judgment with it, then can’t reconstruct why they chose what they chose.
Curious whether your checklist changes when the thing you’re delegating to can’t be accountable.
The mise en place framing for question 3 is the one I’m stealing — that’s the step most people skip.
The line about this working for AI too is where I’d add something. Five of the six transfer cleanly. Question 2 doesn’t, quite.
“Why are we doing this?” works with a person because they carry the why forward into judgment calls you never see. A model takes the why as input and produces better output, but nothing persists. You still own every call it implies.
In coaching I see that gap surface later — leaders who delegated the task and quietly delegated the judgment with it, then can’t reconstruct why they chose what they chose.
Curious whether your checklist changes when the thing you’re delegating to can’t be accountable.
“It will take too long to explain how to do this” might be one of the most expensive sentences an expert can repeatedly say.
I work mostly with deep subject matter experts, and I hear a version of this all the time.
They're right. Initially, it probably IS faster to do it themselves. Explaining “common sense” to someone else can feel frustrating.
So they do it themselves. Again!
Eventually they become the go-to person. Take a number. Get in line. Which can feel like success, but it really isn't.
I've written before about knowledge transfer as an important expert capability, and I think your question “why are we doing this?” is particularly important. Never underestimate the power of the "why".
I would suggest a different take on your questions. They are great questions. But rather than asking them of yourself before you delegate, could you instead take a coaching approach and ask them?
What do you know about this already? How would you go about it?
Why do you think this is important?
What will you need?
What do you think great might look like?
What is a reasonable timeframe to complete this?
What are you worried could go wrong? How might you overcome that?
By asking questions, you turn them from passive task-takers into proactive learners. They are more likely to truly understand the task if you ask them to think through it themselves, with your guidance as a coach.
I would add one concrete example of an acceptable output and name the person who handles exceptions. Otherwise the receiver has to assume both quality and authority.
Delegation gets easier when the handoff includes a definition of done, one example of acceptable output, and the first check-in date. That small upfront cost prevents the manager from becoming the bottleneck.
I spent 27 years on the NYSE floor interviewing seasoned CEOs on TV. I noticed some answered a different question. The question they wanted to answer, versus the one I asked. The audience usually knows.
When I coach senior women today, the failure isn't a confidence problem, it's usually a structural problem: She hasn't pre-built the three sentences.
A no without architecture around it reads as “Oh she’s being difficult” A no with architecture around it reads as graceful and still preserving the relationship.
I call it the “anatomy of a good no.”
The transfer of judgment is what I see most people underestimate. The hard question - the one that decides whether the delegation actually works - is how much context goes along with the task. I've seen this in EdTech too: the briefing that felt thorough rarely was.
Number 4 is the one I only learned through AI. I had a whole page of rules describing the tone I wanted and the drafts kept coming back dull. Then I pinned one message I'd already approved at the bottom of the rules and said match this one. Fixed it immediately, which was annoying after all those rules.