A risk team in a financial services firm had tightened its quality checks on the work of senior colleagues. The process was straightforward. An email set out what the review had found and what needed to change, then a conversation of 30 to 60 minutes agreed the next steps.
The checks themselves were sound. As the review went on, the cases left were the harder ones, and those were the ones most likely to be argued. When the team and I talked about what was going wrong, nobody said the analysis was weak. The trouble started when a senior colleague disagreed.
The pushback came in a few recognisable forms. Some debated the judgement. Some defended their position point by point. Some steered the discussion onto safer ground, usually a technical question they knew well. And some pulled rank.
Why knowing more made it worse
Earlier in the year I’d worked with a team of analysts on a related problem: putting a view across to senior colleagues who had a few minutes to hear it. In both teams, a challenge got answered the way these people had been trained to answer it, with more detail. Every extra paragraph of explanation gave the other person something new to dispute, and made it harder to hear what the speaker actually thought.
With the analysts, we had them give the same message in less and less time. What survived the cut was what they believed mattered. What fell away had been padding, or cover. Done this way, shortening a message tests the quality of the thinking. Nobody lost rigour.
What the Gestalt lens adds
Gestalt practice has a name for the move onto safer ground: deflection. It’s one of the ways people break contact, turning away from what is actually going on between them. This is my lens on it rather than something anyone in the team said to me, but when a senior colleague starts a debate about a methodology point in the middle of a quality finding, the methodology point is rarely what matters to them. It’s easier to argue about than the finding.
The risk colleague who follows them into that debate has joined the deflection. Staying in contact means bringing the conversation back to the point, calmly, and sometimes saying out loud what has just happened: something like “We’ve moved on to the model. Can we go back to the finding?”
What we built
Most of the day was practice in difficult conversations. The team agreed a shared structure for the conversation, so senior colleagues got the same experience whoever they spoke to, and so there was something to return to when a discussion drifted. They worked on a small set of responses for the common patterns: acknowledging a point without conceding it, coming back to the standard, stopping a slide into technical debate, and naming the pattern respectfully when the conversation became avoidant. Then they practised on real, anonymised cases from recent checks, rotating roles, with a follow-up session about eight weeks later.
Afterwards the team told me they felt more confident going into these calls, and that the calls were being handled more consistently from one person to the next.
The other half of the step up
Firms like these promote people for being right. The step up then asks them to be right out loud, to someone who doesn’t want to hear it: a manager telling a former peer their file isn’t good enough, or an adviser telling a long-standing client no. The technical case is usually the easy half.
The other half is that the people around you start talking to you differently. On a programme for newly promoted managers recently, I used a story Julie Zhuo tells in a 2019 Harvard Business Review piece, As Your Team Gets Bigger, Your Leadership Style Has to Adapt. She gave feedback in a design review, asked for questions and got a row of shaking heads, then found out later that the team had disagreed with her. One of her direct reports explained why nobody had said so: “You’re kind of a big deal.”
What changes is mostly quiet. People may tell you less, and tell you later. A question you ask out of curiosity can be heard as an instruction. Old jokes and confidences come across differently, and people watch how fairly you treat the friends you used to work alongside as peers.
My reading of this is that it’s the same deflection running the other way. Nobody argues with the new manager about the finding. They stop bringing the real problem at all, and contact thins out without anyone deciding it should. The new manager is often the last to notice, because the conversations still feel friendly.
Three relationships change at once: with former peers who now report to you, with your new peer group, and with your own manager. Each side has new expectations of the other, and they rarely get said out loud. What helps is asking each of them what they now expect of you, telling them what you expect of them (including what stays the same), and having that conversation early, then going back to it. The questions the managers practised were plain ones:
- “What do you need from me now that’s different?”
- “What would you like to stay as it is?”
- “If I get something wrong, how would you want to tell me?”
The last one matters most to me. It asks for bad news before there is any, and it agrees a way for it to reach you once people have started weighing their words.
If you want to know whether either half is happening in your firm, think of the last quality finding, file review or piece of feedback that came out of its conversation weaker than it went in. Who softened it, and what were they avoiding?

