Most organisations have inclusion values, inclusion training and an inclusion policy. Far fewer have teams where the quietest person routinely disagrees with the most senior one. The gap between the two is not commitment. It’s that inclusion was treated as a message, when it lives or dies as behaviour.
Here is the pattern, and it is remarkably consistent. An organisation commits sincerely to inclusion. There is a statement, a training day, sometimes a network and a calendar of events. Engagement scores on the inclusion questions tick up. And in the daily life of teams — who speaks in meetings, whose ideas get credited, who gets the stretch assignment, who feels able to say “I disagree” — very little moves. The initiative didn’t fail because anyone was insincere. It failed because it operated on what people believe, and inclusion is decided by what people do, in small moments, under time pressure, when nobody is framing it as an inclusion moment.
An inclusive team is, before anything else, one where speaking up is safe — where a half-formed idea, an awkward question or an unwelcome disagreement doesn’t cost you. That safety is not created by declaring it. It is created, or destroyed, by what happens in the three seconds after someone takes a risk: whether the leader gets curious or defensive, whether the interruption is allowed to stand, whether the person who flagged the problem is thanked or remembered. Leaders set this almost single-handedly, and mostly by going first — admitting uncertainty, asking for pushback and visibly rewarding the person who provides it.
Safety is created, or destroyed, in the three seconds after someone takes a risk.
Bias training raises awareness, and awareness fades by Thursday. What endures is changing the decisions where bias actually operates: who gets hired, assessed, promoted, credited and given opportunity. The practical moves are structural rather than heroic — agree the criteria before you meet the candidates; make interviews and reviews consistent enough to compare; when the stretch assignment comes up, ask “who haven’t we considered?” before defaulting to the usual reliable name. None of this requires anyone to feel accused. It requires the team to accept that fair outcomes are a process property, not a character reference.
Teams quietly exclude people through the mechanics of how they work: meetings where the fastest talker wins, decisions made in the channel some people aren’t in, jargon that functions as a border, formats that suit one working style. The inclusive corrections are small and concrete — circulate the question before the meeting so reflectors arrive ready; go around the room rather than letting volume allocate airtime; write decisions down where everyone can see them. Each one is trivial. Together they determine who actually gets to contribute.
The behaviours above become culture when they are enforced sideways, not just downwards — when a peer says “hang on, she wasn’t finished,” or credits the idea to the person it came from, or questions why the list of names looks the way it does. That is allyship in its useful form: small, live interventions with a cost attached. It is also learnable, and teams that practise it in low-stakes moments find it available in high-stakes ones.
What distinguishes teams that sustain this from teams that relapse is that they treat inclusion as a performance variable with evidence attached — who is speaking in meetings, how opportunity and credit are distributed, what the safety questions in the survey say, and whether those numbers move. Measured, inclusion stops being a value the team espouses and becomes a standard the team maintains. And the return is not only moral: teams where everyone can contribute simply have more ideas, catch more errors and hold more talent than teams where a third of the room has learned to stay quiet.