How to Delegate Authority, Not Just Tasks
Most founders who are the bottleneck in their own company delegate extensively and would tell you so honestly. They hand over work continuously. What they retain is the decisions inside the work — the approval, the final call, the veto they have never named — and that retention is why the delegation produces no relief. Moving labour reduces your hours slightly. Moving authority is the only thing that removes you from the critical path. This page is about the difference, and the mechanics of a transfer that actually holds.
What’s inside
Tasks versus authority
The four things you can hand over
The five-part transfer
The fortnight that decides it
How to review without taking it back
What to keep
Where Blomma fits
Tasks versus authority
The distinction is simple and the consequences are large.
Delegating a task: “Can you put together the pricing analysis and bring me a recommendation?” You have moved several hours of work. The decision is still yours, so the work still queues behind your attention, and the person doing it is producing input rather than owning an outcome.
Delegating authority: “You own pricing for the mid-market segment. Decide it. Tell me what you decided and why.” You have moved the decision. The work no longer queues behind you, and the person is now developing judgment because they carry the consequence.
The test for which one you are doing:if they made a call you disagreed with, would it stand?If the honest answer is no, you delegated a task. That question is uncomfortable and it resolves most confusion about whether delegation is real.
The consequence of getting this wrong is not just your calendar. People who produce recommendations rather than decisions do not build judgment, so they stay dependent — which then looks like evidence that they were not ready for authority. That loop is self-confirming and invisible from your seat, because what you observe is a team that does not take initiative.
The four things you can hand over
Worth separating, because founders hand over the first and believe they have handed over all four.
The work.Doing the thing. Easiest, and least valuable on its own.
The decision.Choosing what to do. This is the one that matters, and the one most often retained.
The standard.Judging whether the output is good enough. Retaining this makes you a permanent quality gate, which is a queue.
The relationship.Being the person the customer, partner, or candidate deals with. Frequently overlooked and often the real dependency.
A genuine transfer moves at least the first two, usually the third, and sometimes the fourth. If you have moved only the work, expect no change in your load and no growth in the person.
The five-part transfer
Most failed delegation is a failed handover rather than a failed delegate. Five components, and skipping any one produces the situation where you end up doing it again.
One. Name the decision explicitly.Say out loud what they now decide without you. Vague delegation reverts to the founder under pressure every time, because in the absence of an explicit statement everyone defaults to asking. “You own this” is insufficient; “you decide pricing within these bands, without checking with me” is a transfer.
Two. Transfer the context, in writing.Why it is done this way now, what was tried before, the real constraints, what the board cares about. Most “they got it wrong” outcomes are context you had and did not transmit. This is an hour of writing and it prevents weeks of rework.
Three. Define what good looks like.Specifically enough that they can assess their own output without you. If you cannot articulate it, you have just found the reason you have never been able to hand this over — and defining it is now the work.
Four. Name the boundaries.What would bring them back to you: a spend threshold, a legal question, a customer above a certain size, anything irreversible. Explicit boundaries are what make genuine autonomy inside them possible. Without them, people either over-escalate or over-reach, and both get read as poor judgment.
Five. Set a review point and say what happens at it.“Let’s look at this in six weeks” gives you a legitimate place to put your anxiety. Without one you will check in randomly, which reads as surveillance and quietly undoes the transfer.
Then tell the rest of the company. A handover your team does not know about does not take — people keep coming to you, and you keep answering, and within a month the transfer has silently reversed.
The fortnight that decides it
Everything above is preparation for one moment, and it is not the handover.
About three weeks in, something will go visibly worse than it would have gone with you. You will see it early, you will be right, and stepping in will feel like helping rather than reclaiming. It will be entirely defensible.
If you step in, you have reset the transfer to zero — and worse, you have demonstrated to that person and everyone watching that the authority was conditional. The next handover carries no credibility, because they have now seen what happens under pressure.
So decide the intervention threshold in advance, in writing, while you are calm. Almost always the honest threshold is “irreversible or seriously costly,” and almost always the thing in front of you is neither. If it is below the line, let it happen and hold the review conversation afterwards, which is where the learning is for both of you.
This is the hardest hour of the whole transition and it is where most delegation attempts actually fail — not at the handover, which is easy and feels virtuous, but three weeks later in a moment nobody else recognises as decisive.
How to review without taking it back
You still need visibility. The trick is reviewing the reasoning rather than re-deciding the outcome.
Ask how they decided, not what you would have decided. “Walk me through how you got there” gets you the information you need and leaves the decision theirs. “Why didn’t you do X” is a re-decision wearing a question mark.
Give feedback on the process and the standard, not the specific call. “Next time I’d want the finance input before you commit” is coachable. “That was the wrong price” takes the decision back.
Do it on the agreed cadence, not when your anxiety spikes. Random check-ins are the most common way founders undo their own delegation, and from the other side they read unambiguously as distrust.
And when something goes well, say so specifically and publicly. The fastest way to make a transfer real to the rest of the company is for people to see the person get credit for a decision you did not make.
What to keep
The literal version of this advice produces a founder who has no idea what is happening in their own company.
Keep the genuinely irreversible decisions: market entry, platform commitments with long tails, senior hires and departures, fundraising, anything hard to undo.
Keep the standard at company level — what quality is acceptable, what behaviour is not, who you are willing to hire. Nobody else can hold that line.
Keep direct contact with customers, and enough depth somewhere to smell when something is wrong. Contact, not review authority. This is how founders retain the judgment that makes them better than a hired operator.
And keep the ability to override when a situation genuinely warrants it — rare, explained, and not your default. Used often, it teaches your company that authority is decorative.
Where Blomma fits
The hard part is not understanding any of this. Every founder knows they should delegate. The hard part is that a real transfer is made of specific conversations and one specific act of restraint, and there is nowhere to work either of them out.
Blomma is an always-on AI career coach with no stake in how you run your company. Use it to run the honest test on something you believe you have already delegated — would their call stand if you disagreed? Use it to write the five-part transfer properly, including the standard, which is the part founders discover they cannot articulate. Use it to set your genuine intervention threshold before the three-week moment arrives. And use it in that moment, which is the highest-value use of all: somewhere to take the impulse to step in, late at night, and think it through instead of acting on it.
The second application is the other side of the transfer. The person receiving authority is making calls they have not made before, and the standard outcome is that they do it unsupported and their early mistakes get read as unreadiness. Coaching for them is what makes the delegation compound rather than stall..
Delegation is not about handing over more work. It is about handing over decisions, specified properly, and deciding in advance what you will do when the first one goes badly.