Executive coaching
Coaching vs Mentoring: The Difference, and Which One a Leader Needs Now.
Coaching draws on your experience. Mentoring draws on someone else's. The AI shift is the first leadership transition where nobody's experience covers the job, and that changes which one you need.
When a leader steps into a bigger job, one of the first things they ask for is a mentor, and one of the first things the company offers is a coach. Nobody in that conversation is quite sure whether those are the same request.
The coaching vs mentoring question usually gets answered with a list: coaching is short and mentoring is long, coaching is formal and mentoring is not, one asks and the other tells. All true, and none of it says which one you need.
Here is what does. The difference between coaching and mentoring comes down to whose experience is in the room. A coach works from yours. A mentor works from theirs. For most of the history of coaching and mentoring that was a matter of taste.
It is not any more, because the job most leaders are now being asked to do, leading people and AI agents at once, is one nobody has twenty years of experience doing.
Coaching vs mentoring at a glance.
Strip the vendor lists down and the two differ on eight things. The rest of this article is about which of the eight still hold once AI is in the work.
| What differs | Coaching | Mentoring |
|---|---|---|
| Whose experience is in the room | Yours | The mentor's |
| Who sets the agenda | Agreed in writing at the start | The mentee, meeting by meeting |
| What it is for | A specific goal or decision | Growth over a career |
| How long it runs | Months, with an end date | Years, until it stops |
| How formal it is | A contract, goals and checkpoints | A relationship |
| Who does it | A trained coach, usually external and paid | A more experienced person, usually internal and unpaid |
| What they need to know | How to coach; the seat helps | The job, the industry, the politics |
| How you know it worked | Against the agreed goals, judged by others | Retention and promotion, years later |
Both run on confidentiality, listening, questions before advice, and a trust that takes a few sessions to earn. The line between the two columns is blurrier than the table makes it look, and that is where the useful part starts.
What coaching is.
Coaching is a contract. Someone trained to do it agrees with you on a goal, a time frame and a way of knowing whether you got there, then works with you, mostly by asking, until the time runs out. The coach does not need to have done your job.
They need to get you thinking about it more clearly than you can alone, and to hold you to what you said you would do.
A leader meets coaching in three forms, and they are not the same thing:
- Executive coaching. One to one, external, paid, with a contract. This is what people mean when they say the company bought them a coach.
- Team coaching. The leadership team as a system, worked on together: how it decides, disagrees and delegates. Coaching five people one at a time is not team coaching.
- The manager who coaches. A daily habit, not a program. In Rewired Leadership I call it coaching over directing: the leader stops handing down the answer and asks the question that makes the team weigh the trade-off. Leaders in our programs practice it on their own decisions, because reading about it does not build it.
What mentoring is.
Mentoring is a relationship. Someone with more experience shares what they know with someone who has less, and it runs far longer than coaching, often for years.
There is usually no contract, no fee and no end date. The mentee brings whatever is on their mind, and the mentor answers from what they have seen. It comes one to one, in groups, and in reverse, where the junior person teaches the senior one.
Sponsorship is a different thing: a mentor spends their time on you, a sponsor spends their capital.
The need is real and only partly met. In Deloitte's May 2026 survey of 22,595 Gen Zs and millennials, only 54% of Gen Zs and 60% of millennials said their team could maintain performance if a key expert left tomorrow, and 29% named a lack of time or priority as a barrier to passing knowledge on.
The five differences that actually matter.
Lists of eleven differences between coaching and mentoring are these five restated. Each carries a decision for the person choosing, so the decision sits underneath.
01Whose experience is in the room
The coach draws on yours. The mentor draws on theirs. A coach can work with a surgeon, a founder and a general on the same day because the material is the client's.
A mentor is only as useful as the overlap between what they have done and what you are about to do. So: if nobody has done what you are about to do, coaching is the only one that works.
02Who sets the agenda
A coaching agenda is agreed in writing before the first real session, and a good coach will not let you drift from it without a conversation about why.
A mentee sets the agenda meeting by meeting, from whatever is on their mind that week, which is the charm of it and the limit. So: if a sponsor is paying and wants to see progress, it has to be coaching.
03The clock
Coaching has an end date. The engagements I run are six months, with a session every two weeks, and the end is in the contract because the deadline is part of what makes it work.
Mentoring runs until it stops, which can be a career. So: pick by whether the problem has a date on it. A first ninety days has a date. Becoming a better leader does not.
04What gets measured
Coaching is judged against the goals agreed in month one, and the honest way to judge it is to ask the people around the leader. In my engagements the same five to eight colleagues who described the situation in week two are asked again at the end, so the change is judged by others.
Mentoring is judged, if at all, by whether the mentee is still there years later. So: when you need to prove something changed, it is coaching.
05Who does it, and what it costs
A coach is trained, usually credentialed, usually external and paid. A mentor is usually a volunteer from inside the company and unpaid, which is why mentoring is cheap to start and hard to sustain: it is the first thing pushed aside by urgent work, and most programs quietly die of that.
So: mentoring costs little to launch and a great deal of discipline to keep. Coaching is the reverse.
Where the line blurs, and why that is fine.
The clean version above is not how either practice runs in the wild. Coaches who have done the job get asked for advice, and good mentors coach: SHRM's February 2026 guidance tells mentors to "listen before advising" and to "coach for independence."
So can a coach be a mentor? Many already are, some of the time. The question is whether they tell you which hat is on.
There is a third kind of person in this market, and it is the one I think leaders in a transition need. An operator who has run companies and coaches inside a structure is doing both at once: the contract, the goals and the end date of coaching, with the judgment of someone who has made the call.
The question is not coach or mentor but method or miles, and whether you can get both from a coach who has carried the number. The caveat cuts both ways. A mentor who only tells war stories is not coaching you, however senior.
A coach who has never sat in your seat can only ask, and there are weeks when a leader needs someone who can also say "I would not do that, and here is what it cost me."
Why AI changes the answer.
Every distinction above was written for a world in which the senior person had done the job the junior person was learning. That world is ending at both ends of the ladder at once.
At the top, mentoring is losing its subject
Nobody has twenty years of leading a team made of people and AI agents. Nobody has five. The senior leader's "how I did it" was built in a world where coordinating work, reviewing output and knowing the answer were the job, and those are exactly the things agentic systems now do.
In Rewired Leadership I put it bluntly: AI will not replace leaders, it will expose them, and the habits most likely to be exposed are the ones that made a leader successful. Mentoring transfers experience.
For this transition, a good part of the senior person's experience is the thing to unlearn.
That does not make mentors useless. It narrows what they can give to the things that do not expire: judgment under consequence, ethics, how to handle a board, what fear does to a team and how bad news and mistakes travel through it.
Ask a mentor for those, not for how to run the work, because the work is being redesigned around both of you. The one piece of experience that does transfer is knowing which calls still have to be made by a person, and even that is being learned in real time.
The unlearning runs the other way too: in my coaching work with Gen Z leaders, the digital natives have to be coached to take their most ambiguous problems off-screen and into a real conversation, because judgment is built there and not in asynchronous text.
Neither generation has the experience the other needs. Both need coaching.
At the bottom, mentoring is losing its pipeline
Mentoring assumes an apprentice doing the work next to someone who has done it. In a growing number of companies there is no apprentice.
In the Oliver Wyman Forum and New York Stock Exchange survey of 415 chief executives published in April 2026, 43% said they plan to deprioritize hiring for junior roles in the next year, up from 17% a year earlier.
The companies furthest ahead on AI were the exception: a quarter of them plan to hire more juniors, not fewer (Oliver Wyman Forum and NYSE, April 2026).
Coaching vs mentoring needs an apprentice: what CEOs plan for junior roles
The share of CEOs stepping back from junior hiring more than doubled in a year. A quarter of the companies furthest along with AI plan to hire more juniors, not fewer.
Source: Oliver Wyman Forum and NYSE CEO survey, April 2026, 415 CEOs
Gartner's September 2026 warning to marketing chiefs tempted to cut the first rung is the line to remember: "The biggest risk is mistaking the automation of entry-level tasks for the elimination of the need for entry-level talent."
Its remedy is to give juniors the evaluation of AI output and a seat in strategic decisions earlier in their careers. Look closely at that remedy. Asking a junior person "what would you do with this output, and why" is not mentoring.
It is coaching, applied to work the junior did not produce. The apprenticeship is being replaced, where it is replaced at all, by coached judgment, and the companies that notice will be the ones that still have a bench when they ask where the next leaders come from.
Coaching's price goes up, because the only precedent is yours
If no one's experience covers the job, the only material to work from is the leader's own situation, which is what coaching has always used. That is why the book's model of leadership is built on it at every level.
The rewired CEO becomes the organization's chief executive coach, setting intent and guardrails and coaching managers on the how. The rewired manager stops micromanaging execution and starts coaching judgment.
And the CHRO, whose job changes most of all once agents do the execution, becomes what I call the judgment coach, with one rule: you only coach on the guardrail gap.
When a decision goes wrong, the question is not who failed but where the guardrail failed to cover the situation. That turns the person who made the call into an architect of the system instead of its casualty.
Nothing in that model requires a mentor. All of it requires a leader who can coach.
Can AI be your coach or your mentor?
AI is a good sparring partner, a weak coach and no mentor at all. It can rehearse the hard conversation with you at midnight and give you feedback on the third attempt. It cannot hold you to what you promised, and it has no experience of its own to share.
The evidence is better than most of what gets written about AI coaching. In February 2026 the first randomized trial to compare accredited human coaches with an AI coaching chatbot gave 114 leaders four weekly hours with one of 74 experienced coaches, the same hours with a text chatbot, or a place on a waiting list.
It found "substantial effectiveness in diverse coaching outcomes but only for human coaching, with effect sizes in the mid-to-high range."
Four of the people who dropped out of the AI arm asked for a human coach instead (de Haan, Terblanche and Nowack, HRDI, 2026).
One other study, in California Management Review in November 2025, reported the reverse, but its AI group got more than twice the contact time, so the fair reading is that frequent practice helps and that the accountability still needs a person.
INSEAD's Nadav Klein, writing in May 2026, has the division right: AI is the sparring partner, available at any hour for practice and feedback, and the human coach is the sensei, who provides the accountability, the trust and the context.
Failing to follow through on a chatbot's advice, as he notes, simply feels less consequential. The ground AI is actually taking is mentoring's: in Deloitte's 2026 survey, seven in ten Gen Zs and millennials now seek career advice through AI.
A model can rehearse anything with you. It cannot have been anywhere.
So put it where it belongs: rehearsal before the session, preparation, notes and the nudge between meetings, all of which make the human hour better.
Keep the accountability and the agreed goals with a person, and remember that a system built to be agreeable will agree with you, which is the one thing a leader paying for challenge does not need.
Which one does a leader need? A five-question test.
The test I use in a first conversation is short. Each question points to one of the two, and most leaders in a transition answer yes to more than one.
Coaching vs mentoring examples: five questions that point to one or the other
Is there a decision or a behavior with a date on it?
If yescoaching, with the date in the contract
Has anyone you can reach actually done this exact thing?
If yesmentoring, for as long as their experience holds; if no, coaching is the only one that works
Does a sponsor need to see progress against agreed goals?
If yescoaching, because mentoring has nothing to report against
Are you new to the role, the industry or the politics?
If yesmentoring, for the map
Is the problem the team rather than you?
If yesteam coaching, not five mentors
If nothing above appliesstart with coaching, because the goal has probably not been named yet, and naming it is the first session
Most leaders in a transition answer yes to two of these. That is the case for both, and for finding them in one person.
Three leaders I see most often, and how the test lands for each:
- The CEO or founder leading a shift nobody has run before. Coaching, with someone who has sat in the seat, because there is no mentor for this one.
- The new executive in the first ninety days, leading former peers. Both: a mentor for the politics, a coach for the decision rights they have to let go of.
- The leadership team that has to decide faster together. Team coaching. Five mentors would produce five better individuals and the same slow team.
Running coaching and mentoring inside a company without mislabeling them.
Most organizations need both, and most of the failures I see come from calling one the other.
- Name which one you are buying, in the brief, the budget line and the first conversation with the person receiving it. A mentoring program sold as coaching will be judged by coaching's measures and fail them.
- Give mentors coaching skills. A mentor who asks before telling is worth several who only tell, and asking can be taught in a day.
- Protect senior leaders' time for it, or it will not happen. Unpaid and uncontracted means first to be dropped.
- Make managers coaches. A manager who cannot coach judgment on AI-assisted work cannot manage it, which is why rising leaders need to learn the new habits before the old ones set.
- Measure them differently. Coaching by the goals agreed in month one, a midpoint and an end review, and the same colleagues asked again at the end. Mentoring by retention, promotion and time to productivity, with a date on each, so that neither is judged by activity an agent now does.
- Put AI where it belongs: practice and preparation, never the accountability.
Frequently asked questions.
Whose experience is in the room. A coach draws on yours, through questions, inside a contract with a goal and an end date, and the result is measured.
A mentor draws on their own, informally, for as long as the relationship lasts, and the result is rarely measured at all. The other differences follow from that one.
Yes, and many already are, some of the time. An operator who coaches is doing both: the structure of coaching with the judgment of someone who has made the call. The only rule is that the person should tell you which hat is on at any given moment.
Coaching, for the decision in front of you, because it has a date and a measure and draws on your situation. Mentoring, for the map of a new role, an industry or a board.
For an executive leading a company through the AI shift, coaching does more of the work, because nobody has done that job long enough to mentor on it.
Coaching: a new vice president spends six months with an external coach on which decisions to hand to her team and which to keep, judged at the end by the people who report to her.
Mentoring: a first-time manager meets a senior colleague monthly through a merger to understand its politics. Both examples are illustrative, not client cases.
Usually. Mentors are mostly volunteers from inside the company; coaches are mostly external professionals, paid by the hour or a fixed fee for the engagement.
That is why mentoring is cheap to start and hard to sustain, and why coaching is the reverse: the fee is also the thing that keeps it on the calendar.
Not on the evidence. The 2026 randomized trial that compared accredited coaches with an AI chatbot found substantial gains only for the human coaching.
AI makes a good sparring partner for rehearsal and preparation, and it cannot be a mentor, because it has no experience of its own to share. Keep the accountability with a person.
The short version.
Coaching draws on your experience and mentoring on someone else's, and the AI shift is the first leadership transition where nobody's experience covers the job.
So coaching's value rises, mentoring narrows to the things that do not expire, and the leaders who do best find both in one chair: the method of a coach and the miles of someone who has carried the number. Start by naming the decision in front of you. The rest of the choice follows from it.