How to Help Colleagues at Work Without Burning Out

There is a specific kind of professional who is loved by everyone and going nowhere. They answer every question, take every ad hoc request, cover every gap, and their own work happens after seven in the evening. Their appraisal says they are a great team player and does not say much else, because the things they spend their time on are not the things anyone measures.

This is a solvable problem, and the solution is not to help less. It is to help differently. This guide covers how to help colleagues at work in ways that actually scale, how to teach rather than rescue, how to mentor without it consuming your calendar, how to set boundaries before you resent people, and how to make any of it visible in an appraisal.

What is the most useful way to help a colleague?

The most useful help is the kind that removes the need for the next request. Solving someone's problem helps them once. Teaching them the pattern helps them permanently. Writing down the pattern helps everyone who will hit it after them, including people you will never meet.

That is the whole hierarchy:

Do it for them. Fastest now, guarantees the request returns. Sometimes correct, in a genuine emergency or where the task is genuinely one-off.
Do it with them. Slower now, they can do it next time. The default for anything that will recur.
Write it down. Slowest now, serves everyone forever. Correct as soon as the same question arrives a third time.
Fix the thing that caused the question. The confusing interface, the undocumented process, the missing automation. Highest leverage and almost never done, because it is not what was asked for.

Most helpful people live permanently at level one, which is why they are permanently busy. Moving your default to level two and applying the rule of three for level three changes both your impact and your calendar within a couple of months.

The rule of three

The first time someone asks a question, answer it. The second time, answer it and notice. The third time, stop answering and write it down, then send the link.

This one rule does more for a helpful person's workload than any other single change, and the reason people do not do it is that writing takes fifteen minutes and answering takes three. The arithmetic only works if you count the fourth, fifth and eleventh time you would have answered, which nobody does in the moment.

What to write depends on the question:

A recurring how-do-I: a short runbook with the exact commands or steps.
A recurring why-is-it-like-this: a decision note explaining the reasoning and what was rejected.
A recurring where-is-it: a links page, which sounds trivial and saves an extraordinary amount of collective time.
A recurring judgement call: a checklist that makes the judgement explicit.

Keep them short. A three paragraph document that exists beats a comprehensive one that you never finish.

How do you teach instead of doing?

Sit with the person while they do it, rather than doing it and showing them the result. The distinction sounds trivial and is not: watching someone else work builds recognition, doing it yourself builds capability, and only the second one stops the next request.

A structure that works for a thirty minute session:

Ask them to describe what they think should happen first. This surfaces the actual misunderstanding, which is very often not the one they asked about.

Let them drive. Their machine, their keyboard, their document. If you take over, you have reverted to doing it for them and they will remember nothing.

Narrate your reasoning, not just your steps. "I am checking the log first because in my experience this error usually comes from the upstream job, not from here." The reasoning transfers. The steps do not.

Let them be stuck for thirty seconds before intervening. Uncomfortable and important. The moment of being stuck is where the learning happens, and rescuing them at the first pause removes it.

Finish by asking them to write the note. They document what they just learned. It reinforces it for them, and it creates the artefact so you never do this session again.

That last step is the one people skip and it is the one that makes the whole thing compound.

How do you mentor someone without it taking over your week?

Mentor with structure and a defined scope rather than open-ended availability. Most mentoring relationships collapse because they were never given a shape, so they become an unlimited claim on the mentor's time that both people eventually feel guilty about.

What structure looks like:

Agree a duration. Three months, six sessions, then review. Defined endings make it easy for either person to stop without it being a rejection.

Agree a specific goal. "Get comfortable running the monthly close independently" is mentorable. "Career guidance" is not, and open-ended mentoring drifts into unstructured venting for both parties.

Fixed slots. A recurring thirty minutes fortnightly beats being available whenever. Fixed slots also mean the mentee saves up questions, which makes each session denser.

Make them own the agenda. They come with what they want to cover. If they arrive without an agenda twice, the relationship is not working and you should say so.

Distinguish mentoring from sponsoring. Mentoring is advice, and it is what most people ask for. Sponsorship is spending your own credibility to advocate for someone when they are not in the room. Sponsorship is worth much more and should be given selectively, to people whose work you can genuinely vouch for.

Reverse mentoring is worth mentioning here too. Junior colleagues often hold current knowledge that senior people lack, on tooling, on how the work actually happens day to day, on how the culture feels at the bottom. Asking a junior colleague to teach you something is one of the fastest ways to build a real working relationship, and it costs you nothing but the admission that you do not know it.

How do you give help that is actually useful?

Ask what kind of help is wanted before providing it. A great deal of well-intentioned help fails because it answers a question the person did not ask, or solves a problem they had already decided to live with.

The three kinds people want, and they are rarely interchangeable:

Information. They want to know a fact or a method. Give it directly and briefly.

A second opinion. They have a view and want it tested. Argue with it properly. Agreeing to be pleasant is worthless here.

To be heard. They want to describe a frustrating situation. Solving it immediately is experienced as dismissal.

The fix is one question at the start: "Do you want me to think about this with you, or do you want me to just listen?" It sounds slightly clinical the first time you say it and it prevents an enormous amount of mismatched help.

When you are giving feedback on someone's work specifically, structure it. The situation-behaviour-impact model is simple enough to actually use: describe the situation, describe the specific behaviour, describe the effect it had. "In yesterday's review, when the client asked about timelines, you gave a date without the caveat about the dependency. They have now planned around a date we cannot commit to." That is usable. "You need to be more careful in client meetings" is not.

How do you set boundaries before you resent people?

Set the boundary when you agree to help, not when you have had enough. A limit stated at the start is administrative. The same limit stated in month four, after the resentment has built, is a conflict, and the other person will experience it as a sudden withdrawal rather than as a reasonable line.

Practical phrasings:

On duration. "I can take the weekly report while you are on the audit, so through to the 20th. After that it comes back to you."

On scope. "I will do the analysis section. I am not going to write the narrative, that needs someone who was in the client meetings."

On availability. "I keep Tuesday and Thursday mornings clear for focused work. Anything urgent, message me, anything else will get a reply by the afternoon."

On repetition. "This is the third time we have hit this, so rather than fixing it again I am going to write it up and send you the doc. Come back to me if the doc does not cover it."

On capacity. "I want to say yes and I would be doing it badly. Ask me again after the close."

Two things make boundaries hold. Say them once, calmly, without a justification paragraph, because long justifications invite negotiation. And apply them consistently, because a boundary that moves when someone pushes is not a boundary, it is an opening position.

What if the person you are helping never reciprocates?

Distinguish between someone who cannot reciprocate and someone who will not. Helping a junior colleague who has nothing to offer back is not a bad trade, it is the normal direction of professional generosity, and it comes back from somewhere else eventually. Repeatedly helping a peer who takes and never gives is a different situation and it is worth naming.

Signals you are dealing with a taker rather than someone with less to give:

Requests arrive urgently and always at their convenience.
Your help is never acknowledged to anyone else, including when credit is being distributed.
They are unavailable when you need something, with a reason each time.
The requests grow rather than shrink as you satisfy them.
They ask you to do things rather than to teach them, consistently, including things they could learn.

The response is not confrontation, which rarely works and costs you a relationship you have to maintain anyway. It is quiet recalibration: move your help down the hierarchy from doing to teaching to pointing at documentation, respond on your timeline rather than theirs, and be busy. Most takers move on to a softer target quickly, because the relationship was never about you.

The one case worth escalating is where the pattern is affecting delivery, for example where someone is systematically routing their work to you and taking credit for the output. That is a manager conversation, and it should be framed around workload and attribution rather than around character.

How do you make helping visible without being performative?

Record outcomes, not effort, and do it in the places where records are read: your self-appraisal, your one-on-ones, your CV. Announcing each act of help as it happens is transparent self-promotion and it devalues the thing itself. Documenting the aggregate effect once a quarter is professional.

The translation from effort to outcome:

What you did	How to record it
Answered lots of questions	"Wrote the FAQ that reduced repeat queries on the payments flow by about half"
Helped a colleague constantly	"Mentored two analysts, both now run their reporting cycles independently"
Covered for someone on leave	"Ran the month-end close for two cycles during a colleague's absence with no escalations"
Onboarded new joiners	"Onboarded four joiners, cutting average ramp from six weeks to two using a written runbook"
Helped another team out	"Supported the support team through the billing migration, resolving 30 escalations that would otherwise have reached engineering"

Every entry in the right column names a beneficiary and an effect. That is what makes it creditable. Every entry in the left column describes your input, which is not.

For interviews, the common questions are "tell me about a time you helped a colleague", "how do you support the people around you" and "how are you helping others grow at work". Answer with a capability transfer and a lasting artefact rather than a rescue story. Practising these against follow-up questions is what the Goodspace AI Mock Interview is built for, and structuring the achievement bullets for a CV is what the Goodspace CV maker handles.

Frequently asked questions

How can I help colleagues at work without doing their job? Teach rather than do. Sit with them while they do it themselves, narrate your reasoning rather than just your steps, and ask them to write the note at the end. The extra fifteen minutes is the difference between one instance of help and a permanent dependency.

How do I stop being the person everyone dumps work on? Set boundaries at the point of agreement rather than at the point of exhaustion, move your default from doing to teaching, and write down anything you have explained three times. Also say no clearly at least sometimes, because a person who never declines is treated as capacity rather than as a colleague.

How do I mentor someone without it taking over my calendar? Give it a shape: a defined duration, a specific goal, fixed fortnightly slots, and an agenda the mentee owns. Open-ended mentoring without structure drifts into unlimited availability, which is why most mentoring relationships quietly collapse.

What is the difference between a mentor and a sponsor? A mentor gives you advice in private. A sponsor spends their own credibility advocating for you in rooms you are not in. Sponsorship is significantly more valuable and should be given selectively, only to people whose work you can genuinely vouch for.

How do I say no to a colleague without damaging the relationship? Decline the specific request and offer a route: name someone better placed, offer a smaller version you can honour, or defer to a date you can commit to. State it once without a long justification, because extended justification invites negotiation.

What do I do if a colleague only ever takes and never helps back? Recalibrate quietly rather than confront. Move your help from doing to teaching to pointing at documentation, respond on your timeline rather than theirs, and be genuinely busy. Escalate only if it is affecting delivery or if they are taking credit for your output.

How do I get credited for helping others in an appraisal? Record outcomes rather than hours. Ramp time reduced, tasks a colleague now owns independently, documentation the team uses, cover provided with no escalations. Help described as effort is not creditable. Help described as capability created is.

Is it worth helping people who are more junior than me? Yes. It is the normal direction of professional generosity, it is where teaching skill is developed, and junior colleagues become senior ones with long memories. It is also the fastest way to find out whether you can explain what you know, which is a separate skill from having the knowledge.