Beyond Compliance: What a Workplace Investigation Cannot Do

Why procedurally perfect processes still fail employees, and what HR leaders can do about the gap between a legally defensible file and psychological safety

One person. All that apparatus. No one coming.

The investigation is complete. The report is thorough. The file is legally defensible. By every measure the organization can point to, the process was handled correctly.

And yet. The person who came forward has gone quiet in meetings. The person named in the complaint is disengaged and has retained a lawyer. Two colleagues who were interviewed as witnesses have asked to work from home permanently. The manager has stopped addressing friction on the team entirely, because the last time a concern was "raised," it took eight months and left everyone worse off than before.

A workplace investigation is designed to answer one question: did this conduct occur, and does it meet the threshold for a policy violation? It was never designed to answer the harder one: what do we do now, as an organization of people who still have to work together?

That gap is not a failure of execution. It is a failure of design.

Naming the Injury: Institutional Betrayal and Psychological Safety

There's a term for this in the academic literature: institutional betrayal, coined by psychologist Jennifer Freyd. It describes a specific kind of injury, one that only an institution can inflict, and only on someone who depends on it.

That dependency is the mechanism. An employee doesn't just earn a paycheque from their employer. They draw their professional identity from it, their benefits, their sense of belonging, often the entire trajectory of where their career goes next. Institutional betrayal isn't the original harm itself. It's what the institution does, or fails to do, once that harm has been reported: covering up a known pattern, ignoring an earlier complaint about the same person, retaliating against the person who spoke up, or leaving them exposed in ways the institution had the power to prevent. The investigation that follows a complaint is itself one of those institutional acts, which means it can either compound the betrayal or interrupt it, depending entirely on what the institution does within it.

Freyd's diagnosis holds up. Where her work runs into trouble is in the remedy she proposes: institutional courage, a stance of transparency and accountability. That's a disposition, not a method. You can find organizations with every ounce of that courage, genuinely committed to doing right by the people in front of them, still running processes engineered for an entirely different purpose than actually encountering the human being on the other side of the file.

Good intentions don't solve an architecture problem. Only a different design does that.

The Gap Between Legal Fairness and Psychological Safety

Canadian law sets a real bar here. Since Baker v. Canada, institutions making decisions that affect someone have to clear specific procedural ground: tell the person what's happening, let them respond, keep the decision-maker neutral, give reasons for the outcome. Most workplace investigations clear that bar without much trouble.

Clearing it isn't the same as the process actually working for the people living through it, though. Tom Tyler spent years studying what makes people experience a process as fair, work that reshaped how police departments and courts operate, and what he found keeps coming back to four things: people need to feel they were actually heard, not just interviewed. They need to be treated as a person throughout, not a file. They need to see the same standard applied to everyone, not just to them. And they need some sense that the person deciding their fate actually cared about the outcome.

Run a typical investigation against that list and watch it fail every category while still being completely legal. Notice arrives as a letter addressed to "the Respondent." The chance to respond happens in a recorded interview with someone juggling two dozen other open files. The investigator stays neutral, which in practice often reads as detached. The reasons show up in a report that's translated a person's actual experience into findings and policy conclusions.

The process was fair. You were not seen.

This is institutional betrayal in action: a legally compliant response that still leaves the person more exposed than before they came forward.

Three Registers of Harm: Why Psychological Safety Needs Its Own Pathway

Not all harm asks for the same response. Physical harm calls for safety measures, and in serious cases, the criminal system. Material harm calls for institutional remedies and legal recourse. Both of these have existed long enough to have settled vocabularies and known paths to resolution. Everyone in the room recognizes the playbook.

The third register doesn't have that advantage. Psychosocial harm takes many forms, but the form that formal process handles worst is what happens when an individual stops being treated as a whole person: collapsed into a label, a job title, a gender, an accent, until nothing else about them gets through. Or rebuilt into a version of themselves they don't recognize. Either way, the mechanism is the same. Someone stopped looking closely enough at the human being in front of them, and that's what made the harm possible. Look closely enough, at any point, and the harm gets harder to enact.

Here's the part HR teams often miss. Once this kind of harm has occurred, material remedies are reaching for the wrong register entirely. A finding, a sanction, a policy change cannot restore what was damaged in the space between people. The only thing that can address it is a process built to put the person back together in the eyes of the people around them, to make them feel seen as who they actually are, not who the file made them.

And that responsibility belongs to the institution, regardless of what the other party does. Even if the person who caused the harm won't engage, the organization can still step forward and say: we see what happened, we're treating it as real, and we're changing the conditions of this workplace because of it. That's not a therapeutic intervention dressed up in corporate language. It's intentional institutional design.

What Relational Accountability Looks Like

Relational accountability is more precise than investigation-based accountability, not gentler, but more accurate, because it operates in the register where the actual injury lives.

It begins with naming: what happened, stated plainly, without the euphemism that formal processes tend to produce. It requires that the affected person's experience be heard and held as real, not filed as an allegation, not translated into findings, but present in the room as something that actually happened to a person. It requires a genuine encounter with the human consequences of one's actions. Where the responsible party refuses that encounter, the institution does not wait. It takes responsibility for adjusting the conditions of the shared workplace, because that is within its power regardless of whether the other person shows up.

This means structural change, not just documentation. Changed reporting relationships, required capability-building, or in serious cases, separation. And it means follow-up: returning to ask whether conditions have actually changed, not just whether the file was closed.

None of this requires bilateral participation to have value. But it does require the people inside the process, not just the process itself, to reckon with what actually occurred.

Key Questions

What is institutional betrayal, and how does it show up in HR?

It's a term coined by psychologist Jennifer Freyd, and it's easy to misread as just "the company didn't care." It's more specific than that. The original harm, the harassment, the conflict, the misconduct, is one event. Institutional betrayal is a second, separate event: it's what the institution does, or fails to do, in response. Covering up a known problem. Failing to act on a prior complaint about the same person. Retaliating against the person who came forward. Even a procedurally correct response can trigger it, if the institution's action or inaction in that response leaves the person more exposed or unprotected than they were before they came forward. In an HR context, it's not the original incident that does this damage. It's the institutional act, or failure to act, that follows it. How do you protect psychological safety after a workplace investigation closes?

Closing the file and protecting psychological safety are two separate jobs, and most organizations only budget for the first one. The second requires a deliberate, structured process after the investigation ends, one that names what happened plainly, gives the affected person a genuine experience of being believed, and makes real changes to the working conditions regardless of whether the other party participates. It also requires actually checking back in, at set points like 30, 60, and 90 days, to find out whether anything changed on the ground rather than just on paper. The investigation is over, but the team still won't work together. What now?

The investigation answered a narrower question than the one you're actually facing. It established facts and policy compliance. It didn't, and wasn't designed to, repair the relationship between people who still have to sit in the same meetings. What you need at this stage is a separate, structured process aimed specifically at relational repair, not a reopening of the findings. That usually means a facilitated conversation focused on what people need from each other going forward, paired with real structural change, not just documentation that the issue was addressed. The investigation found in her favour, so why did she quit anyway?

This happens more often than most HR leaders expect, and it isn't a sign that anyone did their job badly. A workplace investigation is built to answer one question: did this happen, and does it violate policy? It was never built to answer the question that actually determines whether someone stays: do I still feel safe and seen here? A finding can be entirely correct and still leave that second question untouched. The complainant doesn't usually leave because the outcome was wrong. She leaves because the process itself never addressed the harm that lives in the relationship, not the file.

A Different Question

The question is not whether conflict will arise. It is what your response to it reveals about your culture. The frameworks described here are not theoretical. They have been developed and tested across organizations of every size and sector, by practitioners who specialize in exactly this work.