Psychological Safety in the Workplace: The Missing Half
- 2 days ago
- 8 min read
I get asked to run workshops for teams that have been described to me, in advance, as having a trust issue. Almost every time, someone in that room had already raised the thing months earlier, to someone more senior, and then watched nothing move. By the time I walk in, the room has learned something durable about how the organization handles inconvenient information, and no trust-building exercise will unteach it.
That is the part of psychological safety we keep skipping.
what psychological safety in the workplace means
The definition everyone works from comes out of Amy Edmondson's 1999 research: a shared belief that the team is safe for interpersonal risk-taking. In practice, that describes a workplace where people ask questions when they do not understand something, admit mistakes, ask for help, disagree with someone more senior, give difficult feedback, and report problems without unreasonable fear of what will happen to them afterwards.
It does not mean comfort, and it does not mean agreement. A psychologically safe team argues. A leader in one will make a decision that half the room opposes and say so plainly. The measure is what people conclude afterwards. If the conclusion is that silence would have been the smarter play, psychological safety was not there, regardless of how pleasant the meeting felt.
the easy half and the hard half
Almost all the attention in this field goes to the seconds before someone speaks. Do they feel safe enough to say it? Much of the psychological safety machinery organizations build is aimed at that one moment.
Far less attention goes to the weeks and months afterwards: who is required to respond, what gets recorded, what happens when the concern is about the manager who would normally receive it, how the person is protected if retaliation becomes a live risk, and what happens when an organization listens attentively and then does nothing at all.
Speaking up is half the system. The response is the other half, and most organizations have built only the first one.
the difference between an idea and a concern
It is easy to encourage people to speak when the subject is a product idea. Someone disagrees with a senior leader in a meeting, the leader asks questions and explains the final call, and nobody is punished for having raised it. That is real, and it counts.
Now take the same employee raising something else. They believe a colleague is being bullied. They question whether a promotion decision was fair. They report their own manager's behaviour. They name a safety problem everyone has quietly agreed not to discuss. They challenge a direction the leadership team has already committed to publicly.
Those situations test something the first one never touches. They test whether the organization has machinery capable of holding difficult information once that information becomes inconvenient, exposes a failure, or creates an obligation to act.
a friendly team is not automatically a safe one
The most common misunderstanding is that psychological safety means everyone gets along. A team will be warm, sociable and genuinely fond of each other while steering around every subject that matters. A manager will be approachable in every ordinary interaction and then react badly the one time someone challenges a decision they have already announced in public. Pleasantness is what an organization looks like when nothing difficult is happening.
The related misunderstanding is that safety means lowering the bar. High-performing systems need access to bad news more than anyone else does. When people are afraid to say I made a mistake, I don't know, I need help, or I think this plan has a serious problem, leaders end up deciding from an increasingly filtered version of reality. That is a performance problem well before it is a wellbeing one.
people are watching, and they are keeping score
There is rarely a single cause. Sometimes it is one manager. More often it is the organization itself, teaching people one case at a time what speaking actually costs.
Employees pay attention to a specific set of events: whether disagreement changes how a person is treated six months later, whether an investigation was run by someone with a stake in the outcome, whether raising the same issue a fourth time produces anything different from the first three, whether an appeal process genuinely tests a decision or politely confirms it, and whether a leader ever faced a consequence.
An organization can run a speak-up campaign every quarter for a decade. If experience keeps teaching people that speaking changes nothing, or makes their own position worse, the poster loses.
four places to look
Individual leadership behaviour matters and will never be sufficient on its own. Leaders can ask better questions, admit when they are wrong, invite challenge and thank people for difficult feedback, and every bit of that is worth doing. The organization still needs machinery around the moments where speaking becomes expensive. Four areas carry most of it.
Leadership accountability. Can leaders themselves be challenged? What happens when one of them shuts down disagreement repeatedly, in front of witnesses, and everybody lets it go? Are the commitments leaders make in response to employee concerns tracked anywhere? Will a serious concern about a leader reach somebody with the independence and the authority to act on it? And when material leadership failure continues, does anything actually follow?
Workplace safeguards. People raise more than ideas. They raise bullying, harassment, retaliation and misconduct, and those need routes that work even when the obvious route is compromised. A safeguard is a live one when it lets a person bypass the individual implicated, when retaliation risk is assessed rather than assumed away, when a conflict of interest disqualifies an investigator, and when a second route exists for the times the first one fails. A reporting channel on its own is a mailbox.
Voice and organizational response. Most large organizations collect employee voice well. Surveys, town halls, listening sessions, open-door policies, forums and suggestion tools all exist and all produce material. What decides whether any of it matters is the obligation on the other side: who owns the response, whether a response is required at all, whether the resulting commitment is tracked, whether a recurring issue escalates on its own rather than waiting for a champion, and whether the organization will be able to show, a year later, what actually happened. Collecting voice without an obligation to respond produces the appearance of participation and none of the organizational accountability that makes participation worth anything.
People decisions. Some of the moments that shape psychological safety most are ordinary administrative ones: a performance rating, a promotion, a pay decision, a change of scope, an exit. Nobody is entitled to agree with every decision. What matters is whether a significant decision will be explained, whether the process will be reviewed by someone without a stake in it, whether a conflict of interest disqualifies that reviewer, and whether a decision found to be wrong gets corrected. When people believe that questioning a decision will worsen their own position, the cost lands on everyone who watched, not only on the person who asked.
what surveys and training carry, and what they don't
Surveys are useful. They show whether people feel able to speak, where teams differ, how that shifts over time, and which corners of the organization need a closer look. What they measure is perception, which is a different kind of evidence from the one organizational accountability runs on. A survey will not establish that every serious concern entered a case system, that an investigation was independent, that nobody experienced retaliation six months later, or that a leader's commitment was ever completed.
Training has the same shape. Managers learn how to respond to challenge, teams practice giving feedback, and awareness genuinely rises. A manager can also sit through a psychological safety workshop on Tuesday and react badly to a challenge on Wednesday, and both of those things can be true. Training describes what should happen. A functioning system determines what has to.
where ISO 45003 comes in
ISO 45003:2021 is an International Standard that provides guidelines for psychological health and safety at work. It sits alongside ISO 45001 in the occupational health and safety family and moved the conversation somewhere important by treating psychosocial risk as a property of work itself, including workload, communication, leadership, relationships and organizational culture, rather than as a question of whether individuals are resilient enough to cope.
Psychosocial risk and psychological safety overlap without being the same thing. ISO 45003 provides a management-system approach to identifying, assessing and managing psychosocial risk. What it does not provide is the kind of case-level, auditable response architecture Vanguard Voices is developing around what happens after a person raises a concern.
from a definition to a standard
If this matters as much as everyone says it does, the useful question becomes what an organization should be required to demonstrate.
A standard asks whether specific mechanisms exist and whether evidence shows them operating: whether significant concerns are traceable, whether leadership commitments are followed through, whether an employee can bypass a conflicted manager, whether retaliation risk is assessed, whether material people decisions are explained and reviewed, whether recurring voice receives an organizational response, and whether serious failures escalate when they need to.
A value statement asks to be believed. A standard asks to be shown.
Read next: What Is a Psychological Safety Standard?
Certification is the question immediately after that. A standard defines the requirement; certification asks whether an organization meets it. Where an organization selects its own evidence, picks its own interviewees, interprets its own failures and decides whether it passed, the exercise produces a logo and very little else.
Independent assessment is what separates the two, and it comes with limits worth stating out loud. No certification can promise that every employee will always feel psychologically safe, and no assessor can guarantee that nobody will ever experience poor leadership or harm. The narrower question is whether an organization has built credible machinery around voice, response and organizational accountability, and whether independent evidence shows that machinery working.
Read next: What Is Psychological Safety Certification?
what to ask inside your own organization
Take one real concern that somebody raised in the last twelve months and follow it all the way through. Find out who was required to respond, what was recorded, how long it was allowed to sit unresolved, what would have happened if the concern had been about that person's own manager, how retaliation risk was assessed, whether the underlying decision was reviewable by anyone independent, and what happened the second and third time the same leadership problem appeared.
Then the harder one: will you be able to show any of that to somebody outside the company?
psychological safety should not run on courage alone
I have spent twenty-five years in corporate life, and I have watched capable people go quiet in the space of a single meeting. It made me sad. Actually sad, not professionally concerned, because most of it was avoidable and nobody in the room intended it.
I am building this while still working inside a large organization. No organization has yet been certified, and version 0.1 is still being tested and challenged. I think that matters, because the Standard should be shaped by the reality of how organizations actually work, not by pretending the difficult parts have already been solved.
Courage will always be part of saying something difficult, and no standard will take that away. What courage should not have to do is carry the entire system by itself.
Organizations already build one half of it: the surveys, the hotlines, the open doors, the listening sessions, the town halls. The other half is somebody being required to receive what comes in, somebody owning the response, the issue staying traceable, protection existing where protection is needed, decisions being reviewable, failure carrying a consequence, and evidence showing what happened.
That is what psychological safety in the workplace actually requires, and it is the half far too many organizations leave unbuilt.
Interested in helping shape the Standard?
sources
Amy C. Edmondson, “Psychological Safety and Learning Behavior in Work Teams,” Administrative Science Quarterly 44, no. 2 (1999): 350–383.
American Psychological Association, “What is psychological safety at work?”
ISO 45003:2021, Occupational health and safety management — Psychological health and safety at work — Guidelines for managing psychosocial risks
