Building Psychological Safety That Actually Works
Authored by: Jenine Saleh Esq.
Psychological safety has become a phrase that gets cited in mission statements, repeated in onboarding decks, and invoked at all-hands meetings without changing very much about how anyone actually behaves. The concept itself is sound. Amy Edmondson’s research on what makes teams effective is some of the best work in organizational behavior of the last two decades.
The problem is what the concept turned into once it left the academy: a kind of cultural good-housekeeping seal that organizations award themselves, usually based on values statements rather than operational practice.
In thirteen years of running legal services for asylum seekers and refugees, I’ve come to think about psychological safety as an engineering problem rather than a cultural one. The work my team does is high-pressure and emotionally heavy. People sit with clients who have survived torture. Cases turn on details that come out only in the fourth or fifth interview, often because a client did not feel safe enough to mention them earlier.
When a junior attorney misses something, or an interpreter signals discomfort with a phrasing, or a case manager flags that a colleague’s tone with a client crossed a line, the cost of those signals not surfacing is not abstract. It shows up in someone’s removal hearing.
That context forced us to be unusually concrete about what produces safety and what merely performs it. Most of what works is not encouragement. It is structure.
Supervision is where psychological safety gets built or doesn’t
The single highest-leverage place to build or destroy psychological safety is the recurring one-on-one between a staff member and the person who supervises them. Most organizations treat these as status updates. They are absolutely not status updates. They are where a staff member learns, over weeks and months, whether it is actually safe to bring up a problem before it has a clean solution attached.
The mechanics matter more than they sound like they should. Showing up on time signals the meeting is real. Asking open questions before reporting questions (“what’s been hard this week” before “where are we on the X case”) signals that the conversation has room for more than throughput. Not pulling out a phone. Not jumping to fix. Sitting with discomfort when something difficult comes up, rather than rushing to a positive frame. None of this is novel. All of it is rare in practice.
The harder work is what supervisors do when a staff member reports something that cannot easily be fixed or that the supervisor would prefer not to hear. That moment is the test. If the response is defensive, dismissive, or quick to redirect responsibility back to the person raising the issue, every other person in the organization who hears about it learns not to bring things up. And people talk. The norms get set by what happened the last time someone said something hard.
The response to exclusionary behavior is the policy
Organizations also tend to have a written policy on inappropriate or exclusionary conduct. Almost all of those policies are functionally fiction, because the policy itself is not what gets tested. What gets tested is what happens in the ten minutes after a staff member tells a manager that a colleague made a comment that landed badly.
The response needs to be prompt, specific, and visible enough that the person who raised the concern can see it was taken seriously, without making the situation worse for them. That last clause does a lot of work. A response that turns into a lengthy formal process, that requires the person who raised the issue to repeat their account three times to different audiences, that pulls them into meetings with the colleague they reported, is often a response that punishes reporting. The mechanics of how seriously something gets taken should not be experienced as a tax on the person who already absorbed the harm.
This is where power dynamics across race, language, disability, caregiving status, and seniority show up most starkly. The patterns are predictable. The person who gets interrupted in meetings is usually not the senior white attorney. The person whose accommodation request keeps getting deferred is usually not the partner-track hire. If the response to those patterns is “we’ll keep an eye on it,” what people learn is that the eye is decorative.
Attribution shapes who feels like an owner
Who gets credit for work in your organization is a question of psychological safety, even though it usually gets discussed in other terms. When a case settles favorably and the press release names the partner law firm but not the staff attorney who wrote the brief, the staff attorney learns something about how their contribution is valued. When a strategic decision becomes “leadership’s call” and the pushback that shaped it is invisible in the framing, the person who pushed back learns whether their voice actually moved anything.
Fair crediting is not about making people feel good. It is about whether people feel like owners of the work or hands attached to it. Owners surface problems earlier. Owners take risks. Owners stay. Hands keep their head down and execute, and that is rational behavior on their part if the credit structure has taught them that visibility flows in one direction.
Concretely, this means writing case summaries, board reports, and external communications with attribution that reflects who actually did the work. It means letting junior staff present to board members rather than always staffing those meetings with senior people. It means naming sources of ideas in real time when decisions get made. The cost of doing this is small and falls almost entirely on senior people whose names are already known. The benefit compounds.
Inclusive decision-making is where most organizations stop short
The hardest part of operationalizing psychological safety, and the part most organizations never reach, is meaningful inclusion of underrepresented staff in actual strategic decisions. Not in the focus group convened to validate a decision already made. Not on the committee whose recommendations get filed politely. In the room where the budget is set, where the hiring slate is built, where the priorities for next year are decided.
The practical obstacle is usually time and seniority. Strategic decisions get made by the people whose calendars and titles allow them to be in those conversations, and the demographics of who has those calendars and titles skew in predictable ways. The remedy is not just inviting more people. It is restructuring how those decisions get made so that input from staff outside the senior tier is treated as load-bearing, not as a courtesy.
Written input that gets read before the meeting. Specific questions that ask for substantive views, not reactions. Decisions that visibly change based on what was raised.
What changes when the structure is real
When these practices are in place, what shifts is not morale, although morale shifts too. What really shifts is the quality of the work. People surface problems earlier, when they are still fixable. Junior staff catch things senior staff would have missed, because junior staff are closer to the day-to-day. Concerns about a colleague’s behavior get raised before they metastasize. Strategic decisions get better because more of the relevant information is in the room.
Inclusion that is operationalized is inclusion that actually functions. Everything else is decoration, and most people in the organization, especially the ones with the most reason to pay attention, can tell the difference.
Author Bio: Jenine Saleh Esq., Executive Director, Global Health Conscious NFP
