Why delegation feels unsafe

By Sonja Cilliers and Maryke Swarts, PMRI · 12 September 2026

Every managing partner knows the pattern, and most are living it. The work that should sit with an associate sits with a director, because the director has learnt, over years, that it comes back wrong, or late, or needing so much correction that doing it herself would have been faster. Delegation, in theory, frees the senior person for the work only she can do: the client relationship, the strategy, the opinion that carries her name. In practice, in most firms, it feels unsafe, and so it does not happen. The director stays late. The associate goes home at six, wondering why nothing of consequence ever reaches his desk.

The same pattern runs through every corner of the profession. A general counsel with a team of eight still drafts the board memorandum herself. A senior advocate briefs junior counsel and then rewrites the heads of argument from the first paragraph. A sole practitioner hires a candidate attorney and finds, six months later, that the candidate is doing the filing and the practitioner is still doing everything else. In each case the senior person is carrying work that someone else could do, and in each case the explanation they give is about the other person.

The usual explanations are all about the juniors: they are not ready, they do not care enough, this generation is different. Some of that is occasionally true. The explanation that survives examination, though, is about how the senior brain is processing the risk, and it is worth understanding because it is the one thing in the picture the senior person can actually change.

What the senior brain is doing

Legal training builds a brain that scans for what can go wrong. This is the professional asset: the ability to see the clause that will fail, the fact that will be disputed, the deadline that will be missed, the argument the other side has not yet thought of. Over twenty years of practice this vigilance becomes the default setting, running continuously in the background, and it does not switch off when the task changes from reviewing a contract to deciding who should draft it.

So the decision to delegate is processed as a risk assessment, and the vigilant brain produces a vivid, specific picture of the downside: the draft that goes to the client with the error in it, the call from the other side, the name on the letterhead, the conversation with the managing partner. The picture arrives complete, in colour, with sound. The upside, a junior who learns and a director with an afternoon free for the work that actually needs her, is abstract and delayed and has no picture at all. The brain weights vivid and immediate over abstract and delayed every time; this is one of the most reliable findings in the study of decision-making. Delegation loses the assessment before it has begun, and the director keeps the work, feeling responsible rather than aware that a bias has decided for her.

There is a second layer. The senior brain is also, usually, tired. A prefrontal cortex under sustained load has less capacity for the slow, deliberate weighing that would correct the bias. So the decision defaults to the fast one, and the fast one is always “I will do it myself”.

What it costs

The cost lands in three places, and none of them appears on a timesheet. The director’s own decision quality declines, because she is carrying a volume of work that keeps her prefrontal cortex in a state of permanent overload, and the opinion she signs at eleven at night is the one the client will rely on. The junior does not develop, because the only work that develops a junior is work with real consequences, and that is precisely the work being withheld; a candidate attorney who spends two years on tasks that cannot go wrong learns nothing about how to handle the ones that can. And the firm’s capacity stops growing at the number of hours its senior people can bill, which is the ceiling every small and mid-sized firm eventually hits, and the point at which the partners start talking about a merger.

The method

The method is to change what the senior brain is assessing rather than to argue with it. Delegation feels unsafe because the picture is of the junior’s work going out unchecked. Build the checkpoint into the delegation itself and the picture changes. “Draft this and bring it to me on Wednesday with your three main uncertainties marked” is a very different risk from “draft this and send it”. The vigilant brain now sees a controlled process with a review built in, and the review is scheduled for a time when the director will have capacity for it, rather than at midnight the night before it is due.

The instruction to mark uncertainties does something further. It gives the junior explicit permission to say “I was not sure about this”, which most juniors will otherwise hide. Amy Edmondson’s study of work teams, published in 1999, found that the teams that learn fastest are the ones in which members feel safe to admit error and ask for help, and that this safety is set largely by the leader’s behaviour; the paper is the origin of the term psychological safety, which has since been rather overused, but the underlying finding holds. A junior who is asked for his uncertainties will bring them. A junior who is expected to have none will bury them in the draft, where the vigilant senior will later find them and conclude that delegation does not work.

Then correct differently. The instinct of the vigilant brain is to fix the draft, which is fast, teaches nothing, and produces the next draft with the same errors, which confirms the belief. Coaching the junior to find the error takes ten minutes longer once and removes the error permanently. The question is “what does clause 4 need to do, and does it do it?” rather than a rewritten clause 4. A director who does this for a month finds that the second month’s drafts need noticeably less of her time, and the third month’s need less again.

Finally, keep a record. Vigilance discards evidence of things that went well, because things that go well contain no threat and the filter has no use for them. A director who writes down, for one month, every task she delegated and what actually happened usually finds that the vivid picture of disaster occurred zero times, that two drafts needed real work, and that most came back adequate or better. The belief driving her overwork had no evidence behind it. Seeing that on paper changes the risk assessment in a way that no amount of reassurance from colleagues can.

The director in the opening paragraph is still responsible for the work that carries her name. She is simply no longer doing all of it, and the associate down the corridor has started staying until seven, which he does not seem to mind.

PMRI’s recorded session The Leadership Dilemma in Law, available with its workbook, goes through this mechanism in full: why the vigilant brain resists delegation, how to structure it so that the risk assessment comes out differently, and how to correct in a way that develops people. For firms, the same material is the core of the leadership development and coaching work we do with partners and directors, shaped around the time the firm has, whether that is a morning, a full day or two days, and described further on our training for law firms page. The programmes are developed jointly by a behavioural specialist and an advocate with 27 years in practice. If you would like to talk about what this looks like in your firm, start a conversation with us. There is no charge for the first one.