TL;DR
- Core decision: The interviewer is the most important choice in the exit process, and the line manager is the worst default.
- When doing nothing is right: Tiny teams with no HR contact, departing employees with no feedback to give, or separations already negotiated through counsel.
- What has to be true: The interviewer must be independent of the reporting line, must hold confidentiality against the manager's wish to know, and must convert the conversation into a written record someone else can act on.
- How the options split: Inside HR (business partner, skip-level), genuinely outside (neutral third party, external provider), or no interviewer at all (a written form).
- Decision rule: If the manager has been part of why the person is leaving, or if the manager has any role in the reference that follows, the manager doesn't run the conversation.
- Outcome to expect: A signal that the leaver wouldn't give in person, a manager who gets feedback without identifying detail, and an organisation that can defend the choice on paper.
The conversation that decides whether the conversation is worth having
A HR business partner, Priya, opens her calendar on a Tuesday and sees a thirty-minute slot booked by a senior engineer who gave notice the day before. The engineer's manager, a director named Tom, has already sent two emails asking to be in the room. Tom's argument is straightforward: he knows the engineer's work, he has sat next to this person for four years, and he will get more out of the conversation than a stranger could. The HR partner has a counter-argument she can feel but not, yet, articulate: the engineer asked for HR to run it.
Priya has forty minutes before the slot starts, a stack of three other exits behind this one, and a director who will read the notes from whichever conversation happens. She knows what the right answer feels like. She doesn't yet know how to defend it to Tom, to her CHRO, or to the engineer herself if she is asked why the manager isn't in the room.
The decision she is making isn't who is most pleasant to sit with. It isn't who will fill the form fastest. It isn't even, strictly, who will be kindest to the leaver. The real issue isn't whether the manager should run the exit interview. The real issue is whether the conversation will produce anything true, and whether the organisation can act on what it hears.
Best tools for Employee Retention
When you genuinely do not need to act on this yet
Not every organisation has the problem this article solves, and pretending otherwise wastes the reader's time. There are four honest stages, and the first one is "do nothing".
You run a tight shop with five leavers a year and the manager already hears it. Picture a design studio of fourteen people, a founder-led agency where the founder knows everyone's partner's name, or a family-owned manufacturer whose workforce has barely changed in a decade. The owner knows who is leaving, why, and what to do about it before the resignation letter is typed. The conversation that matters is the one that already happened over coffee on Tuesday, not the one scheduled for Friday. In a team this small, an "exit interview" run by anyone other than the owner is theatre. The cost of building a separate channel exceeds the value of what it would catch. Skip this article. The risk that matters here is the legal one around final pay and accrued leave, which is a payroll matter, not a data-collection matter, and the data-collection question doesn't even arise.
You have friction but no pattern. Picture a scaling company with around thirty staff, exits rising to one a month, and a HR coordinator who sends out a Google form and files the responses under the leaver's name. Managers grumble that they never see what is in the form. The CHRO suspects the form is being answered carefully because the leaver knows the manager will eventually ask. There's real loss here, and a better interviewer would fix it, but the system isn't failing the business yet. Treat this as the stage to design a new approach before the next round of exits turns friction into damage. The signal you're missing is the signal that would have told you whether to invest in this.
You have a real pattern and you can't name it. The team that built your best product has lost three people in six months, and the departures have all gone to direct competitors. The pattern is invisible because each manager ran their own exit conversation and the notes sit in inboxes no one has opened since. The CHRO has the volume but not the signal. This is the stage where wrong interviewer choices have already cost a year of organisational learning. A new approach now is the cheapest option left, because the next three exits will either confirm the pattern or contradict it, and the firm needs an interviewer who can hear either answer cleanly.
You're about to be asked the hardest question. A senior leaver, a protected disclosure on file, a works council in a country where consultation is mandatory, or a unionised site where the exit conversation has a procedural shape you can't wave away. At this stage the interviewer choice is no longer a process decision. It's a legal and political one, and you need both a named interviewer and a paper trail that explains why. This article is for the stages before this one. Past this stage, take local advice.
Five questions this reader asks themselves at 11pm
"Can we afford to take this away from the line manager?" The reader's worry isn't about money. It's about the relationship cost, the quiet damage to a manager who feels cut out of their own team. The reason behind the question is the reader's memory of the last time a manager was sidelined from an HR process, and how that manager behaved for the next six months. The answer is yes, you can afford it, if you replace the conversation with something rather than remove it. If the manager hears nothing at all, the manager will quietly rebuild the channel the next time someone gives notice, and the rebuilt channel will be worse than what you removed. What goes wrong is the gap with nothing in it.
"Is the leaver going to be honest with HR either?" The reasoning behind this one is the reader's experience of HR as a function that reports upward and protects the firm. Probably not completely. The leaver doesn't know what the HR person will do with what they hear. They're running the same risk calculus the manager would have triggered, just with a different person and a different set of cues. What goes wrong is the assumption that HR is automatically safe because HR is not the manager. HR has to earn safety, in the setup, before the conversation. A HR person who is unknown to the leaver, who shares a weekly meeting with the line manager, or who is visibly junior to the leaver hasn't earned it.
"Do we actually use what we collect?" The reasoning is the reader's honest knowledge of what happens to the folder after the meeting. If the notes go into a shared drive and are read at the next quarterly review, the interviewer doesn't matter, because nothing downstream of the notes matters either. What goes wrong is building a serious process around a habit that doesn't exist. A serious interviewer is more expensive than an unserious one, and the cost only earns itself back if someone reads the memo, believes it, and changes something.
"What does the manager do with the feedback afterwards?" The reasoning is the reader's memory of feedback given badly. This is the question nobody answers well. Some feedback the manager needs, because it's about workload, tooling, or process, and the manager is the only one who can fix it. Some feedback the manager should never see, because it names them, and the manager would act on the naming rather than the issue. The interviewer is also the gatekeeper of that line, and most setups skip the role entirely. The notes go back to the manager and the manager reads them as a verdict rather than as a pattern.
"Could the same conversation, run by the wrong person, do real harm?" The reasoning is a senior leaver the reader once saw handled badly. Yes. A senior leaver in a small office who says something identifying about a colleague, then watches that colleague receive a "developmental" note three weeks later, won't give the next exit conversation to your organisation. The interviewer is also a confidentiality function. What goes wrong is treating the role as a notetaker. The notetaker produces a transcript. The interviewer produces a defensible decision about what to record and what to throw away, and that decision is the entire job.
Three honest categories the approaches split into
Inside HR, separate from the manager. A HR business partner, a L&D lead, a skip-level leader two rungs above the leaver. The defining feature is that the interviewer is in the company but not in the chain of command that's being commented on. When this works, it works because the leaver recognises a face they have seen in other rooms, so the conversation feels less like an interrogation and more like a handoff between two people who will both still be at the firm next week. The HR person has context the leaver doesn't have to explain from scratch, so the conversation can get to the real issue faster. When it fails, it fails in two specific ways. The HR partner is junior to the manager and the leaver senses it, so the leaver shapes their language around what the HR partner could plausibly action, which strips the difficult content out. Or the HR partner is too close to the leadership team and the leaver suspects notes will travel upward overnight, in which case the leaver says the safe version of the story. The example: a People team at a mid-stage SaaS company where the HRBP sits in the same weekly standup as the line managers and runs exit interviews on a Friday afternoon. The HRBP has good faith and knows the work. The leaver knows the HRBP and the manager sit in the same room on Monday morning. The signal is weak.
Genuinely outside the reporting line. A neutral third party inside the company who isn't in HR at all, such as a head of compliance, an internal audit lead, or a respected senior who has no stake in the leaver's outcome. Or an external provider, hired for the conversation and nothing else. The defining feature is that the interviewer can't be rewarded or punished by the line manager for what gets written down, because the interviewer isn't in the manager's chain at all. When this works, it works because the leaver can say the unsayable. The leaver knows, or senses, that this person will not be in the next performance review meeting with the manager, so the leaver has no reason to perform. When it fails, it fails when the external provider's notes are treated as gospel by leadership without anyone checking what the manager's side of the story would have been. A clean memo from a clean stranger reads as authoritative, but it carries the same single-perspective risk as the manager's own notes, dressed up in better writing. The example: a former general counsel who runs three or four exit conversations a month for a 400-person firm, sends a one-page memo to the CHRO only, and never meets the line manager. The signal is strong because the channel is genuinely independent. The weakness is that no one in the firm can interrogate the memo, because the author is outside the firm's accountability.
No interviewer at all. A written form, sent by HR, returned by the leaver, possibly with a promise of anonymity above a small team size. The defining feature is that there's no human whose body language, accent, or rank can suppress the answer. The form doesn't react. The form doesn't follow up on the difficult thing the leaver half-said. The form doesn't make the leaver feel they have to perform a version of themselves for a stranger. When it works, it works for leavers who find a face-to-face exit conversation performative and would rather type, for leavers who are writing in their second language and think faster than they speak, and for high-volume scenarios where pattern recognition matters more than depth. When it fails, it fails for everyone who needs the question to be asked in a different way, who would never have written the worst thing on a form but would have said it if asked once, in the right order, with a pause after the second question. The form never pauses. The example: a written-only process at a 60-person fintech where leaver volume is high enough that pattern recognition matters more than depth, and the form is reviewed by a HRBP who never meets the leaver in person. The form is fast and the data is structured. The HRBP has no way to ask the question they didn't think to put on the form.
Five diagnostic questions to self-assess against
Did the manager contribute to the exit? The question is a careful one. Not "is the manager blameless", which is unknowable from inside the firm. Instead, look at the resignation conversation. Did the manager push back, agree, or seem unsurprised? A manager who agrees immediately is a signal, because it suggests the manager expected this and didn't fight for the leaver to stay. A manager who fights it is also a signal, because it suggests the manager wants to win an argument the leaver has already finished. To check this in your own organisation, ask the manager, in writing, what they think drove the decision. Compare it to the leaver's account later, when the leaver's account exists. The gap between the two accounts is your risk.
Does the manager write the reference, or influence it? If yes, the manager can't also run the exit interview. The reference is a separate power over the leaver's next two years, and the leaver knows it, and the leaver will shape their answers around it whether you ask them to or not. To check this, look at your last three references. Who signed them, and who fed into the wording. If the manager fed into the wording, the manager is the gatekeeper of the leaver's next move, and the leaver will be polite to the manager in the exit conversation.
Is the HR contact senior enough to carry what they hear? A HR coordinator who has been in role for six months won't hold a difficult conversation with a fifteen-year veteran of the firm. The veteran will dominate the room without meaning to, because rank is a habit. To check this, ask the HR contact to describe, on paper, how they would handle a leaver who says the manager bullied a colleague. If the answer is vague, or relies on "I'd escalate", the channel isn't safe enough. The escalation route has to be named before the conversation, not after it.
Does the leaver know, in advance, what happens to what they say? If the only documentation of confidentiality is the interviewer's instinct, the leaver is being asked to trust a stranger's character. That is not a process. To check this, write down, today, what you would tell a leaver who asks who sees the notes. If you can't write it down in one paragraph, the process isn't ready for a senior leaver, because the senior leaver will ask, and the answer you give will be the answer you have to defend later.
Has the manager been told, in advance, that they won't see identifying feedback? If not, the manager will ask, and the answer you give in the corridor after the conversation will be worse than the answer you'd give on paper. To check this, send the manager an email today describing the rule. If the email feels impossible to send, the rule has not been earned yet, and the first exit conversation under the rule will be the one that breaks it.
Five People Who Could Run It, Reviewed
The line manager
The manager sits closest to the work, knows the leaver, and is the only person who can connect what the leaver says to the decisions that actually shaped the leaver's last year. For a clean, amicable departure where the leaver is moving for reasons outside the team, the manager can extract detail that a stranger would miss: which project was rewarding, which one was not, which stakeholder was draining.
The weakness is the one that matters. A leaver who has something difficult to say won't say it to the person they're describing. The manager can't know, in advance, whether they're part of why the person is leaving, and the leaver can't know whether the manager has already decided to make the reference easy or hard. Even if the manager is blameless, the asymmetry is permanent. Add to that the fact that the manager has every reason to steer the conversation toward what is comfortable for them, and that the manager's own performance review may depend on the leaver saying nothing unkind, and the case for the manager running it collapses outside a narrow set of clean exits. Where this fails hardest is a senior leaver with a complex story: the manager runs the conversation, gets the leaver's polite version, writes up notes that flatter the team, and the organisation never learns the thing it most needed to learn.
An HR business partner
The HRBP is inside the company, knows the context, and is the conventional answer in most mid-sized firms. The HRBP can also close the loop: feed themes back into compensation reviews, manager coaching, and the next hiring round, because the HRBP sits in the room where those decisions happen.
The weakness is two-layered. The first is seniority. A HRBP who is junior to the manager can't hold the room when the leaver is a veteran. The second is proximity. In most organisations the HRBP and the line manager share a weekly meeting, sit on the same project reviews, and review each other's work. The leaver doesn't always see this, but they often sense it. The leaver calibrates. Where this fails hardest is a HRBP who is also the recruitment partner for the manager's team. The leaver suspects, correctly, that the same person who is hiring the leaver's replacement is taking the notes that will guide that hire. The notes will say what the hire needs to hear, not what the leaver needs to say.
A skip-level leader
Two rungs above the leaver. The skip-level is far enough from the daily work to feel safe, and senior enough to carry weight. In a well-functioning organisation, the skip-level is also the person who, next quarter, will be reviewing the manager's performance, which gives the conversation a quiet gravitational pull.
The weakness is also the strength, turned around. The skip-level has power over the manager's career. That power is what makes the leaver trust the channel. It's also what makes the manager distrust it. A skip-level who hears something difficult about a manager is now in possession of a fact they can't unknow, and which they can't share without breaking the confidentiality they promised the leaver. The skip-level can also become the new bottleneck for exits in that team: every resignation becomes a quiet favour the manager feels they have to repay. Where this fails hardest is a small leadership team, where the skip-level is also the manager's manager, and where the leaver's feedback about the manager travels one floor up in a way that the leaver didn't consent to.
A neutral third party outside the reporting line
A head of compliance, an internal audit lead, a respected senior engineer who has no stake in the leaver's outcome, a former partner now working on internal projects. The defining feature is that this person has no line to the manager, no skin in the leaver's next role, and no reason to write up the conversation in a way that flatters anyone in the room.
The weakness is practical, not principled. This person is doing a job that's not their job, and they will eventually drift back to their own work. Continuity is fragile. Training is informal. And the leaver doesn't know what this person does, which can feel odd, even if it's exactly what makes the channel work. Where this fails hardest is when the third party becomes the de facto owner of "people listening" inside the firm, which is a role they didn't ask for and the firm has not paid them for. The role expands. The compensation doesn't. The third party burns out, and the channel goes quiet.
An external provider
Hired for the conversation. No employment relationship with the leaver, no future in the firm, no reference to write. The provider's only product is the conversation and the memo that comes out of it.
The weakness is the most subtle of the five. An external provider's notes are often treated as more authoritative than they are. Leadership reads a clean memo from a clean stranger and assumes the truth of it, without the triangulation that an internal process would have built in. A provider can also flatten signal: the same one-page memo for every leaver, because the template drives the conversation. Where this fails hardest is when the provider is also the firm's outplacement partner, the same firm that will help the leaver find their next role. The leaver suspects, sometimes correctly, that badmouthing the manager in this conversation will shorten the list of warm intros the outplacement team is willing to make.
The Decision Table
| Situation | Scale | Setup | Primary Pain | Recommended Starting Point |
|---|---|---|---|---|
| Owner-led team of under 15, low leaver volume | Tiny | No HR contact at all | Owner already knows | Keep exit conversation with the owner or a nominated senior, document the reason in one line |
| HR coordinator in post, exits rising to one a month | Small | HR exists, no process | Form being answered carefully | Move to HR business partner with skip-level backup, manager excluded by default |
| Mid-sized firm with monthly exits and an active People team | Mid | HRBP exists and is credible | Notes never reach the manager in usable form | HRBP runs the conversation, manager receives a feedback memo without identifiers |
| Senior leaver, manager visibly part of the exit story | Any | HR exists | Leaver will not be honest with the manager | Neutral third party inside the firm, skip-level only if no closer option |
| Unionised or works-council site, consultation rights engaged | Any | HR and employee-representative bodies | Procedural shape of the conversation | External provider plus internal co-interviewer, written protocol agreed in advance |
| Leaver signals protected disclosure or formal grievance | Any | HR aware | Legal exposure on what is asked and how it is recorded | External lawyer-led conversation only, treat the meeting as a record |
| Pattern of exits from one team, manager suspected but unproven | Any | HR and leadership aware | Need signal without burning the channel | External provider first, internalise the role once the pattern is read |
| Scaling firm with no HR at all | Small | No HR | Owner cannot also be interviewer | External provider quarterly, internal person trained once volume justifies it |
Making the Choice Defensible
The decision has to survive two challenges: the manager who wants to be in the room, and the CHRO who wants to know why the manager isn't. Both challenges get answered the same way, by writing the choice down before the conversation happens, and by pointing to that note when the question is asked. A defensible choice is one that names the interviewer, names the reason, names what the manager will and won't be told, and names the date on which this was decided. Without those four elements, the choice will be re-litigated at the next difficult exit.
The single biggest reason managers push back isn't ego. It's the legitimate feeling of being cut out of a conversation about their own team. The defence against that feeling isn't exclusion. It's structure. The manager gets feedback, in writing, in a form that doesn't identify the leaver, within a timeframe that lets them act on it. The manager doesn't get to ask who said what. The HR partner holds that line. If the HR partner can't hold that line, the HR partner is the wrong interviewer.
| Scenario | Interviewer | Reason |
|---|---|---|
| Clean exit, leaver moving for life reasons, manager blameless | Manager, with HR present as a witness | Depth of knowledge outweighs risk; HR holds the written record |
| Manager part of the reason, or unclear | HR business partner or skip-level | Independence from the reporting line is the safety criterion |
| Senior leaver, small leadership team, manager suspected | Neutral third party inside the firm or external provider | Confidentiality cannot survive inside the chain of command |
| Legal exposure, works council, protected disclosure flagged | External provider or counsel | The conversation is now a record, not a chat |
| No HR exists at all, owner is also the manager | External provider only | Owner cannot interview themselves |
What the Manager Should and Should Not Be Told Afterwards
Feedback that travels back to the manager has to do three things at once: give the manager something to act on, protect the leaver from identification, and protect the next leaver from learning that their words travelled. The trade-off between those three is real and it doesn't resolve itself.
The manager should be told the themes. Compensation, workload, hiring, recognition, tooling, decision-making. The manager should be told the specific things the leaver flagged, in language that doesn't name the leaver, in language that doesn't date the feedback to a specific exit, and in language that doesn't let a small-team manager work out who said it. The manager should be told soon enough that the feedback is still useful, and late enough that the leaver has finished their notice and left cleanly.
The manager should not be told who said what. The manager should not be told the specific phrasing the leaver used. The manager should not be given the notes. The manager should not be told in a side conversation that they were "fine" or "not fine". A HR partner who hedges in a corridor is the single most common reason the next exit interview produces nothing.
| What the manager receives | What the manager does not receive | Why |
|---|---|---|
| Themes in plain language | Quotes or specific phrasing | Quotes are identifying in small teams |
| Patterns across multiple exits where the pattern exists | Single-incident verdicts | A pattern is actionable, a verdict is not |
| A follow-up conversation with HR about what to do | The raw notes | Notes are personal data and disclosable |
| Time to act before the next exit in that team | A deadline tied to the leaver's leaving date | Pressure invites identification |
What to Put in Writing
Written records are what turn a good decision into a defensible one. They're also the part of this process that almost every team skips, because the conversation felt like enough. It isn't. Without records, the choice of interviewer is a memory, the feedback to the manager is hearsay, and the next leaver starts from zero.
| Artefact | Who owns it | When it is written | What it prevents |
|---|---|---|---|
| Note naming the interviewer and the reason | HR lead | Before the conversation is booked | Re-litigating the choice at the next exit |
| Exit conversation memo, anonymised | Interviewer | Within a working day of the conversation | Notes getting lost, or being edited under pressure later |
| Feedback memo to the line manager, themes only | HR lead | Within a week of the conversation | The manager asking who said what |
| Record of what the leaver was told about confidentiality, in writing | HR lead | Sent before the conversation | Disputes about what was promised |
| Process document naming who runs exit conversations by category | HR lead | Reviewed annually | New managers re-asking the same question |
| Retention schedule for exit notes | HR lead, with legal review | On creation of the process | Over-retention becoming its own problem |
| List of which themes have been escalated, and to whom | HR lead | At each escalation | Themes being raised, forgotten, raised again |
Two artefacts above deserve a note. The first is the confidentiality note to the leaver. It sounds bureaucratic. It's the only way to defend yourself when a leaver later claims they were promised anonymity they didn't receive. The second is the retention schedule. Exit notes are personal data about a living individual. They're usually disclosable to that individual on request. A schedule that says when notes are deleted isn't optional housekeeping. Confirm the position locally before setting the schedule, because the rules around personal data of this kind differ.
Questions to Ask Before You Commit
Independence: Who, in the firm, has no line to the manager and no stake in the leaver's next role? If the answer is "no one", that's your starting problem. A bad answer sounds like "the CHRO is independent of everyone", which is rarely true in a flat leadership team.
Continuity: Who runs the conversation when the named interviewer is on holiday, has left, or has moved teams? A bad answer is a single name with no backup.
Loop closure: What does the manager receive, in what form, within what timeframe? A bad answer is "we'll figure that out afterwards".
Confidentiality mechanism: What do you tell the leaver, in writing, about who sees the notes? A bad answer is "we'll keep it confidential", which isn't a mechanism.
Triangulation: How do you check a single exit conversation against what the manager's side of the story would have been? A bad answer is that you don't.
Documentation: Where are the notes, who owns them, how are they disposed of? A bad answer is that the notes live in someone's inbox.
Pattern recognition: Who reads across exit notes, and how often? A bad answer is that no one does.
Provider integrity: If you use an external provider, who owns the notes they produce, and what happens to them if you stop using the provider? A bad answer is that the provider owns them, which makes your pattern blind to your own turnover.
Legal shape: What is the firm's position on personal data, works council consultation, and disclosure on request in the jurisdictions you operate in? A bad answer is that nobody has checked. Confirm locally.
The cost of getting this wrong
The cost that appears on an invoice is small. The cost that doesn't appear is the one that lasts. A wrong interviewer choice produces notes that flatter the manager, themes that flatter the leadership team, and a firm that believes its own retention story because nobody has heard a different one. The pattern shows up eighteen months later as a wave of exits that nobody saw coming, and the post-mortem names "the market", or "compensation", or "a difficult year", because the early signal was never captured.
The deeper cost is reputational, and it travels. A senior leaver who feels the exit conversation was a performance won't recommend your firm to the next person they know who is looking. A junior leaver who watches a senior leaver be managed through the exit won't give you their honest answer when their turn comes. So the second exit conversation is harder than the first, and the third harder than the second, and by the time the CHRO notices, the channel has gone quiet without anyone turning it off. The question to ask, before you decide, isn't "what does this cost to run". It's "what does it cost us to be the firm that nobody will say the true thing to".
When you are ready to go further
If the patterns above describe your firm and you're choosing between a HR business partner, an internal neutral, and an external provider, the next move is a side-by-side comparison of how each option would handle three of your last exits, with the names redacted. That comparison is the kind of work HROpsLab does, as a review publication, without supplying software, payroll, or advice. We read what providers publish, sit in on what they sell, and write up where the marketing overreaches the reality. If you want a starting point for that comparison, the editorial team can point you at the pieces that line up with your stage and your scale. No pitch, no follow-up sequence, no obligation.
Frequently Asked Questions
Should the manager ever run the exit interview?
Yes, in a narrow band of exits where the manager is clearly not part of the reason the person is leaving, the leaver is comfortable with the manager, and a HR person is present as a witness and as the keeper of the written record. Outside that band, the manager running the conversation is the single most common reason the conversation produces nothing useful. The asymmetry doesn't go away because the manager is well-liked.
What do we do in a small company with no HR?
The choice is between the owner, who can't interview themselves if they're the manager, and an external provider hired for the conversation. The written record still needs an owner. The owner keeps it. The provider runs the conversation. Confirm the position on personal data locally before any notes are filed, because the rules differ by jurisdiction and the notes are personal data about a living individual.
Should we use an external provider?
External providers earn their fee in three specific situations: when the leaver is senior and the story is complex, when the manager is plausibly part of the reason for leaving, and when the firm needs the conversation to be defensible on paper. Outside those three, an internal HR business partner or a neutral third party will do the same job at lower cost and with closer follow-through.
How do we protect anonymity in a small team?
You don't, fully. In a five-person team the manager will often work out who said what, and the only real protection isn't telling the manager at all. Anonymity in a small team is a flag, not a feature. If the team is that small, the interviewer should not be the manager, and the feedback to the manager should be about themes across multiple exits, not about this one.
What do we tell the manager afterwards?
The themes, in plain language, soon enough to be useful, late enough that the leaver has left, in a form that doesn't identify the leaver. Not the quotes. Not the notes. Not a corridor hedge. The manager doesn't need to know who said what to act on what was said, and the moment the manager learns who said it, the next exit conversation is shorter and less honest.
Should we record the conversation?
No. Recording creates a personal-data artefact that the leaver can request, that you have to store, and that doesn't improve the quality of what is said. Written notes, taken by the interviewer, shared with the leaver for correction, are enough. Recording also changes the leaver's risk calculus in real time, because a leaver who knows the conversation is being recorded will perform for the record.
Who sees the notes?
The interviewer, the HR lead, and anyone the leaver has been told in writing will see them. The line manager doesn't, unless the leaver has explicitly agreed to be identified. Notes are personal data about a living individual and are usually disclosable to that individual on request, which is the constraint that shapes everything else. Confirm the position locally before setting any retention rule, because the rules differ by jurisdiction and the constraint can shift what the notes contain.
HROpsLab is a review publication. We don't sell software, payroll, or advice. We write what we find.
