The Principal Investigator as Academic Mentor: What the Role Actually Requires
The PI mentoring role is real but rarely written down. This article sets out what it requires beyond supervision, how to judge a PI's mentoring capacity before you commit, and what you can and cannot expect.

The Principal Investigator as Academic Mentor: What the Role Actually Requires
A PI mentor develops the person, not just the project. Supervision and line management cover the work and the contract. Mentoring covers your development, and you should assess a PI's capacity for it before you join.
You are about to join a lab, or you are already in one, and you are trying to work out what your principal investigator owes you as a mentor. The person who signs off on experiments and approves leave is not the same as the person who is supposed to help you become a scientist. The mentoring role is real, and it is rarely written down. Here is what it requires, how it differs from supervision and management, how to judge a PI's mentoring capacity before you commit, what you can and cannot expect, how to raise problems, and how the relationship shifts as you move from PhD to postdoc to faculty.
What does a principal investigator do as a mentor that is not research supervision?
Research supervision covers the work: which experiments to run, how to design them, when to stop a line of inquiry, how to write the paper. Mentoring covers the person doing the work. A PI mentor helps you build a scientific identity, a network, and a set of judgment calls that outlast any single project. That help often looks like a conversation in a hallway rather than a meeting with an agenda.
One concrete mentoring act is helping you choose problems that fit your skills and your career goals, not just the lab's funding. A PI who knows you can see that you are good at longitudinal data analysis but miserable at bench work, and can steer you toward a collaboration that uses the first skill. Supervision would be assigning you to the bench anyway because the grant says so.
Another is sponsorship. Mentoring includes putting your name forward for things you cannot easily access yourself: a talk at a small invited workshop, a review assignment, a seat on a committee, an introduction to an editor. A PI mentor spends social capital on you. You can be supervised perfectly well by someone who never mentions your name in a room you are not in.
Mentoring also covers the unwritten rules. How to handle a referee who is being rude. When to push back on a co-author. How to say no to a request from a senior colleague without burning the relationship. How to negotiate a start-up package. These are not research skills, and they are rarely taught in courses. A PI mentor makes them explicit, or at least lets you watch them being practiced.
Career planning is part of the role. A PI mentor asks what you want to be doing in five years and then works backward. That might mean encouraging you to teach, even if teaching does not help the lab's next paper, because you want a faculty job at a teaching-intensive institution. It might mean writing a fellowship application with you that takes three weeks of the PI's time and brings no money to the lab.
Emotional and professional support during setbacks is another piece. Experiments fail. Papers get rejected. A PI mentor helps you separate a bad result from a bad scientist. A PI should not be your only source of support, and they are not trained to treat anxiety or depression. But a mentor notices when a setback has knocked you sideways and adjusts expectations for a few weeks. Supervision is often blind to that.
Finally, mentoring includes honest feedback about your weaknesses. A PI mentor tells you that your seminar was hard to follow, that your writing is dense, that you interrupt people in meetings. Supervision may never surface those things because they do not affect the data. Mentoring surfaces them because they affect your career.
How is PI mentoring different from line management and day-to-day supervision?
Line management is about the terms of work: hours, leave, pay, contracts, performance reviews, grievances. In many institutions the PI is also your line manager, which is why the roles get confused. The skills are different. A good line manager runs a fair rota and handles a leave request promptly. A good mentor helps you decide whether to apply for a fellowship. You can have one without the other.
Day-to-day supervision is about the task. It is the postdoc or senior student who shows you how to run the assay, checks your notebook, and tells you the incubator is broken. The PI may do some of this in a small lab, but in a larger group the PI is often too far from the bench to supervise daily work. That distance is normal and not neglect. The PI's mentoring role operates at a different timescale: months and years, not days and weeks.
Supervision is typically directive: the supervisor says what to do next. Mentoring is more often Socratic. The mentor asks what you think you should do next, and why. That difference matters when you are stuck. A mentor helps you learn how to choose tasks, while a supervisor gives you a task. If your PI only ever gives instructions, you are being supervised, not mentored, and you may need to ask for the other thing explicitly.
Line management has formal power. The PI can approve or deny leave, sign off on a thesis, or decide whether to renew a contract. Mentoring has influence but not authority. A mentor can advise you to turn down a collaboration; they cannot forbid it. When the two roles sit in the same person, as they usually do, you need to know which hat is on. A conversation about your career is mentoring. A conversation about your contract is line management. Mixing them up creates confusion, especially when the advice is unwelcome.
There is also a difference in who owns the agenda. Supervision agendas are set by the project: the next experiment, the next paper, the next grant deadline. Mentoring agendas are set by you: your skills, your network, your next move. A PI who only ever runs project meetings is not mentoring. A PI who schedules a separate conversation every few months about your development is. If your PI does not schedule it, you can.
One more difference: mentoring can outlast the employment relationship. A line manager stops managing you when you leave. A mentor may still be writing you references and taking your calls a decade later. That is why the mentoring relationship deserves more deliberate attention than it often gets. It is one of the few parts of academic life that is genuinely long-term.
What should you check about a PI's mentoring capacity before you join?
Start with the people who have already left. Ask former PhD students and postdocs, not current ones, what the PI was like as a mentor. Current lab members may be cautious about criticizing someone who still controls their future. Former members have less to lose. Ask specific questions: Did you meet regularly? Did the PI read your drafts? Did they introduce you to people? What happened when you disagreed? Would you work with them again? The answers will tell you more than any website.
Look at where former mentees went. A PI with a strong mentoring record will have a track record of people moving into jobs that fit them, not just jobs that fit the lab. That pattern is not proof of good mentoring, because people succeed despite bad mentoring too. But a complete absence of former students in independent positions is a signal worth investigating. Ask the PI directly why that might be.
Ask about the structure of mentoring. Does the PI have regular one-to-one meetings? How often? Are they about the project or about you? Is there a written plan for your development, or an annual review that covers more than publication output? Some institutions require these things; many do not. A PI who can describe their mentoring structure clearly is more likely to actually do it.
Check the PI's workload. A PI with three major grants, two editorial roles, and a department headship may be an excellent mentor in principle and unavailable in practice. Ask how many people they currently supervise and how much time they spend with each. Ask what happens when they are traveling. If the answer is that you will be supervised by a senior postdoc, that can work well, but you should know it before you join.
Ask about their approach to authorship, data ownership, and conference travel. These are the areas where mentoring failures show up first. A PI who is vague about authorship is likely to be vague about other things. A PI who never sends students to conferences is not sponsoring them. You are looking for specifics, not assurances.
Talk to the PI about a time a student struggled. Ask how they handled it. The answer will tell you whether they see mentoring as part of the job or as an interruption. You can also ask what they expect from you as a mentee. Good mentors have thought about that question. Poor ones have not.
Consider your own needs. If you are the first person in your family to go to university, you may need more help with the hidden curriculum than someone whose parents are academics. If you are an international student, you may need help with visas and cultural norms. If you have a disability or caring responsibilities, you need a PI who will engage with those things rather than treat them as inconveniences. Ask directly. A PI who is defensive about such questions is telling you something.
What can a mentee reasonably expect from a PI mentor, and what is outside the role?
You can reasonably expect regular access. That does not mean an open door at all hours. It means a predictable meeting, at a frequency you both agree on, where your development is on the agenda. Once a month is a common minimum for PhD students; postdocs may need less. If the meeting keeps getting cancelled, that is a mentoring problem, not a scheduling problem.
You can expect honest feedback on your work and your plans. Not constant praise, not constant criticism. Feedback that is specific enough to act on. If you ask whether you are ready for a particular job market, you should get an answer that reflects the PI's actual view, not a polite evasion.
You can expect sponsorship within reason. A PI mentor should introduce you to people, nominate you for appropriate opportunities, and write references that are detailed and on time. They cannot create opportunities that do not exist, and they cannot make a hiring committee choose you. But they can make sure you are visible.
You can expect confidentiality about personal matters. If you tell your PI that you are struggling with your mental health, that should not become lab gossip. There are limits: if your safety or someone else's is at risk, the PI may need to act. A good mentor will tell you those limits in advance.
You can expect help navigating institutional processes: annual reviews, thesis committees, conflict resolution, complaints. The PI does not have to know every rule, but they should know where to send you. That is part of the role.
What is outside the role? A PI is not your therapist. They can listen and signpost, but they are not trained to treat anxiety or depression. If you need clinical support, ask for it. A PI is not your career placement service. They can advise and introduce, but they cannot guarantee a job. A PI is not your friend, though some mentoring relationships become friendly. The power differential remains, and pretending otherwise can make it harder to raise problems. A PI is not responsible for your entire wellbeing. They have a duty of care, but you also need a life outside the lab.
It is also outside the role to expect a PI to fix a dysfunctional institution. If the department has no mental health support, no ombudsman, and no clear grievance procedure, your PI may not be able to solve that. They can advocate, but they are one person. Knowing the difference between what your PI owes you and what the system owes you will save you a lot of frustration.
How do you raise a mentoring problem with your PI, and when do you escalate?
Raise it early and specifically. Vague complaints ("I don't feel supported") are hard to act on. Specific ones ("Our one-to-one meetings have been cancelled four times in a row, and I need help choosing my next project") are easier. Say what you need, not just what is wrong. Ask for a meeting with that purpose stated in advance, so the PI knows it is not a routine project update.
Choose the right moment. Not immediately after a grant rejection, not in front of the group. A scheduled conversation, in private, with an agenda you have sent ahead, gives both of you a chance to prepare. If you are nervous, write down the three points you want to make and take them with you.
Name the pattern, not the person. "I've noticed that feedback on drafts has been coming after deadlines" is easier to hear than "you never read my work on time." The first invites a fix. The second invites a defense. You are trying to solve a problem, not win an argument.
If the first conversation does not work, try again with a different framing. Maybe the PI did not realize the problem was serious. Maybe they are overwhelmed. A second attempt, a month later, with a concrete request ("Can we put a recurring meeting in the calendar for the next six months?") is reasonable. Keep a record of what you asked for and what happened. Not as evidence for a future complaint, but because memory is unreliable and patterns are easier to see in writing.
Escalate when the problem is affecting your work, your health, or your rights, and the PI has not responded to a direct request. Escalation routes vary by institution. Common ones include a departmental director of graduate studies, a head of department, a faculty ombudsman, a union representative, or a human resources business partner. Your institution's handbook will list them. If you cannot find them, ask a senior student or a trusted staff member.
Escalate immediately if the problem is harassment, discrimination, bullying, or a safety issue. Those are not mentoring problems to be solved with a conversation. They are formal matters. Use the formal route. A good mentor will support you through that process, but they may also be the subject of the complaint, which is why you need someone else.
Escalating does not always mean making a formal complaint. It can mean asking a third party to mediate a conversation, or asking for a change of supervisor. Those are big steps, and they have consequences. Talk them through with someone you trust before you take them. But do not let fear of the PI's reaction keep you in a situation that is damaging you. The power differential is real, and institutions have procedures precisely because of it.
How does the PI mentoring relationship change from PhD through postdoc to faculty?
During the PhD, the PI mentor is often the primary scientific parent. They teach you how to think, how to write, how to present, how to be a member of a scientific community. The relationship is close and hierarchical. You need more direction at the start and more independence at the end. A good PI mentor gradually hands over control: first the experiments, then the project direction, then the writing. If they are still directing every step in year four, that is a problem.
As a PhD student, you should expect the PI to help you build a thesis committee that actually works, not just a list of names. That means people who will read your work and give honest feedback, not just people the PI likes. The PI mentor should also help you think about what comes after the PhD, even if you are not sure you want to stay in academia. That conversation should start in year two, not year four.
As a postdoc, the relationship shifts. You are more independent, and the PI mentor's role becomes more about sponsorship and less about instruction. You should expect the PI to treat you as a colleague in training, not as a senior student. That means involving you in grant writing, inviting you to meetings with collaborators, and supporting your own fellowship applications. It also means giving you space to develop your own ideas, even if they are not the lab's priorities.
The postdoc is often when mentoring gaps become visible. A PI who was a good PhD supervisor may be a poor postdoc mentor because the skills are different. Postdocs need help with independence, not with technique. If your PI cannot let go, you may need to find a co-mentor or a mentoring committee. Many institutions have formal postdoc mentoring programs; use them.
As faculty, the relationship becomes peer-like, though rarely equal. Your former PI may become a collaborator, a referee, or a competitor. A good mentor helps you navigate that transition: how to negotiate your first grant, how to handle a conflict of interest, how to say no to your former advisor without damaging the relationship. They may also become a source of emotional support when you realize that faculty life is not what you expected.
At each stage, the mentoring relationship should become more reciprocal. As a faculty member, you may find yourself mentoring your former PI in some areas, such as new methods or institutional politics. That is healthy. It means the relationship has matured. If it has not, if your PI still treats you as a student a decade later, you may need to set boundaries.
The constant across all stages is that mentoring requires maintenance. It does not happen automatically. You have to ask for it, shape it, and sometimes end it. A PI mentor who is good at one stage may not be good at another. That is not a failure; it is a reason to build a network of mentors rather than relying on one person for everything.
Frequently Asked Questions
What is the difference between a PI mentor and a PI supervisor?
A supervisor directs the research: which experiments to run, how to analyze data, when to publish. A mentor develops the person: career planning, sponsorship, feedback on skills, and help navigating the unwritten rules of academia. The same person often does both, but the roles are distinct.
How can I tell if a PI will be a good mentor before I join their lab?
Talk to former lab members, not current ones. Ask specific questions about meeting frequency, feedback on drafts, introductions to collaborators, and what happened when someone disagreed with the PI. Look at where former mentees ended up. Ask the PI directly how they structure mentoring and what they expect from you.
What can I reasonably expect from my PI as a mentor?
Regular access through scheduled meetings, honest and specific feedback, sponsorship within reason, confidentiality about personal matters, and help navigating institutional processes. A PI is not your therapist, your career placement service, or your friend, and they cannot fix a dysfunctional institution.
How do I raise a mentoring problem with my PI?
Ask for a private meeting with the purpose stated in advance. Be specific about what you need, not just what is wrong. Name the pattern rather than the person. If the first conversation does not work, try again with a concrete request. Keep a record of what you asked for and what happened.
When should I escalate a mentoring problem beyond my PI?
Escalate when the problem is affecting your work, health, or rights and the PI has not responded to a direct request. Escalate immediately if the problem is harassment, discrimination, bullying, or a safety issue. Use formal routes such as a director of graduate studies, ombudsman, union, or human resources.
How does the PI mentoring relationship change from PhD to postdoc to faculty?
In the PhD, the PI is often the primary scientific parent, providing close direction that should gradually decrease. As a postdoc, the PI becomes more of a sponsor and less of an instructor. As faculty, the relationship becomes peer-like, though rarely equal, and may involve collaboration or competition.
Is it okay to have more than one mentor?
Yes. A single PI cannot provide everything. Many people build a network of mentors: one for scientific direction, one for career strategy, one for personal support. Your PI may be part of that network, but they do not have to be all of it.
What if my PI is a good scientist but a poor mentor?
That combination is common. You may need to find mentoring elsewhere: a committee member, a senior colleague, or a formal mentoring program. You can also ask your PI for specific things you need, such as introductions or feedback, rather than expecting them to intuit it. If the gap is damaging, consider changing labs or escalating.
The notes on how to write the paper cover the next question, so this page does not repeat them.
The notes on a fellowship application cover the next question, so this page does not repeat them.
The notes on Papers get rejected cover the next question, so this page does not repeat them.
The notes on former PhD students and postdocs cover the next question, so this page does not repeat them.