Development conversations that start with the coach

SAM needs to classify behaviours to be useful, but that doesn’t mean it has to judge coaches. We’ve taken on feedback from coaches and coach developers, so it surfaces moments for consideration, offers rationale without criticism, and gives coaches space to reflect first, so the conversation that follows is supportive, not defensive.

Classification ≠ condemnation

Any tool that labels behaviours, questioning, instruction, sarcasm, praise, is, by definition, making a call on what it sees and hears. We don’t pretend otherwise. The point is what happens next. Do those labels arrive as a verdict, or as a starting point for reflection?

We designed SAM so the output is an invitation, not a score.

What we learned (the hard way)

In early pilots we produced tidy “observation reports” mapped to detailed behavioural frameworks. They looked rigorous, and they landed badly. Too many categories, too much overlap, and a tone that felt like judgement, even when the intent was developmental. Several coaches told us the report was overwhelming; coach developers said it made delicate conversations harder, not easier.

So we changed course.

From reports to moments

Instead of delivering a top-down report, through the dashboard, SAM highlights interactive moments aligned to simple lenses (e.g., questioning vs instruction, speaking time, relationship cues). There are extra layers where the type of question or feedback, for example, are also identified, but each highlight comes with:

• A prompt, not a verdict “This moment may be worth a look” rather than “This was good/bad.”
• Rationale without criticism “Flagged for sarcasm, this can read as a negative relationship cue in some contexts.”
• Context first Acknowledge that tone, history and intent matter. The same quip can build rapport with one player and bruise another.

Coach Developers working across Professional Clubs, National Governing Bodies and Higher Education have all identified how hard it can be to start a conversation tactfully when asking a coach what made them decide on this approach, or to interact with an athlete in that way.  If you can get the coach to start the conversation, you’re really onto something.

SAM’s prompts do exactly this by allowing the coach to arrive there themselves by gaining access to their session in a timely manner, so the coach developer can follow with a supportive discussion rather than “shining a light” on them.

Example: the “sarcasm” flag

If SAM detects sarcasm, it marks the clip as a potential negative relationship moment. That isn’t a moral judgement; it’s a heads-up. In review, the coach can say, “In this dressing room, with this player, that tone is part of our rapport.” Equally, they might watch it back and think, “Fair point, on a bad day, that could land poorly.” The value is the reflection, not the label.

Why non-judgemental design helps coaches

• Quicker language awareness Seeing patterns (“I used that word 100+ times”; “I defaulted to instructions late on”) accelerates self-awareness without shame.
• From reflection-on-action to reflection-in-action Because insights arrive quickly, coaches start adjusting during sessions, pivoting from a statement to a question, choosing silence over a filler phrase.
• Towards “pre-flection” Over time, coaches anticipate their own drift and catch themselves before the moment, an intentional pause that changes what’s said next.
• Psychological safety Starting with the coach’s view reduces defensiveness; the debrief becomes a joint enquiry, not a trial.

Why it helps coach developers

• Better openings You can begin with: “Which two clips do you want to explore?” rather than “Here’s what I noticed.”
• Shared reference points Timestamped clips anchor the conversation in evidence, not memory.
• Less admin, more mentoring With analysis automated, your time shifts to meaning, not marking.
• Scalable consistency Common lenses across coaches produce comparable artefacts without imposing a rigid, top-down curriculum.

A simple non-judgemental workflow

1. Coach intention (one line) Before the session, the coach notes an aim (e.g., “Finish with questions, not instructions”).
2. Capture & auto-analysis Upload the recording; SAM returns timestamped moments and light rationale.
3. Coach-first reflection Coach chooses two clips and adds a quick note or voice reflection. No grading.
4. Supportive debrief Developer asks, “What did you intend? What actually happened? What will you try next?” Use SAM’s rationale to contextualise, not condemn.
5. Track gently Trend badges (“highest questioning rate so far”, “directive talk reduced at the end”) show if changes stick, without turning growth into a scoreboard.

Guardrails that keep the tone right

• Language matters: in the UI, we avoid judgment words; we use consider, notice, explore.
• Privacy by default: private coach channels and controlled sharing protect trust.
• Context prompts: every flagged moment nudges a “could the context make this appropriate?” check.
• Fewer, better categories: reduce overlap; keep lenses legible in real life.
• Clip-first reviews: resist the temptation to lead with totals; start with moments.

What non-judgemental looks like in practice

• A coach watches a flagged exchange and decides to phone a player; they agree small adjustments and clearer check-ins.
• Another coach notices how directive talk creeps in late; next session, they plan prompts for the final phase, and stick to them.
• A developer avoids a confrontational opener by asking the coach to pick the first clip; the conversation stays curious, not combative.

In each case, SAM didn’t decide who was right. It helped the right conversation happen sooner, and with more care.

Non-judgemental doesn’t mean neutral. It means useful: surfacing the moments that matter, offering just enough rationale to think with, and letting people exercise professional judgement, where coaching happens.

‍