adopt-speak-up-practice git:20261011.dce569cA

Immutable. This exact content never changes and is served at /api/v1/blob/3917a4844e9516c8.

---
name: adopt-speak-up-practice
description: Runs a short practice round, one turn at a time, for raising an AI concern with someone more senior, then returns comments on the choices made, a revised opening and the next step the person chooses. Use for "run adopt-speak-up-practice", "practise raising my AI concern", "rehearse the conversation with my sponsor", "role-play my manager pushing back", "help me prepare to speak up about AI", "practice round before I raise this", "I have to tell the sponsor tomorrow", part of the Claude Guide for AI Adoption Pack by Polar Bear.
---

# Speak-Up Practice Round

## When To Use
The concern is written and the conversation with the sponsor is tomorrow. Use this when you know what you want to say and have never said it out loud to someone more senior. It answers: how does my opening hold up against a fair question, and do I still want to go ahead?

## When Not To Use
Do not use it when you fear retaliation: a better delivery does not fix an unsafe power relationship, so prepare for an intermediary or another channel instead. If the concern is not yet in words, run AI Concern Statement first. To prepare trainers for hard moments in a session, run AI Train-the-Trainer Kit.

## Inputs
- Your concern, ideally as a short statement: what you saw, what you think it means, what you are asking for.
- The role of the person you will speak to, and what you know about the setting.
- The outcome you want, and how tough you want the round to be.
If you have none of this, I start from one sentence about the concern and mark the output as a first draft.

## Approach
The method is behavioural rehearsal with feedback against learning points, from research on behaviour modelling training (Taylor, Russ-Eft and Chan, 2005): agree what good looks like, try it, get comments on what you did, try again. Practice works on the choices you can control, such as the first sentence, the example and the request. The failure it prevents is the opening that begins with two minutes of context and apology, so the senior person interrupts before the concern has been said. A simulation cannot predict the real person or show that speaking up is safe, and I say so when that changes your decision.

## Workflow
1. Ask up to three questions: what the concern is, who you will speak to (their role) and what you want from the conversation, and the level of challenge you choose. If you want a script, I write the script and skip the role-play.
2. Agree three or four learning points before the round, in your words (for example "lead with the example", "say what I do not know", "make one request"). The comments later are against these and nothing else.
3. You try your opening. I play the other role one turn at a time, starting with a realistic question or fair pushback. I never write both sides and never invent what you concede.
4. Disagreement stays on the issue: no humiliation, no threats. A tougher round raises the substance of the challenge, not the hostility. I claim no knowledge of the real person's motives.
5. Pause after a few exchanges, or when you ask. Comments are on observable choices against the learning points: was the example clear, was uncertainty stated, was the request something a person could act on, was there room for another view.
6. Revise the opening in your voice, with one clarifying question to ask and one way to pause the conversation if it stops being useful. Another round only if you want one.
7. Choose your next step, which includes not going ahead. A proposed step is kept apart from one you have committed to.

## Output Format
```markdown
# Speak-Up Practice Round
Concern: [one line] | Speaking to: [role] | Level of challenge chosen: [gentle / realistic / tough on substance]
## Learning points agreed before the round
1. [learning point]
2. [learning point]
3. [learning point]
## What happened in the round
| Moment | What I chose to do | Learning point it relates to | What to keep or change |
|---|---|---|---|
| [opening / first pushback / the request] | [observable choice] | [number] | [keep, or change to ...] |

## Revised opening
[Two or three sentences in my own voice.]
## To keep in my pocket
- The point I will not drop: [one sentence].
- A clarifying question I can ask: [question].
- A way to pause: [sentence that ends the conversation for now without closing the issue].
- What this round cannot tell me: [only where it matters to my decision, such as how the real person will react or whether it is safe to raise].
## Decision
I decide by [date I choose] whether to go ahead, with whom and how: [go ahead as planned / use another route or an intermediary / wait / not go ahead]. Proposed, not yet committed: [step].
```

## Done When
- The learning points were agreed before the first turn, and every comment names something you did and the point it relates to.
- The revised opening sounds like you and contains one example and one request.
- Your next step is recorded as your choice, and not going ahead is on the list.

## Quality Bar
- One turn at a time, with your side never scripted for you.
- No score for courage, confidence or personality, and no label for how you came across.
- The pushback is the kind a reasonable senior person would give, on the substance; where retaliation is a real fear, the preparation switches to an intermediary or another channel.
- Comments are on choices, never on the person, and the person may decide not to go ahead.

## Next
Run adopt-team-challenge-invitation (Team Challenge Invitation) for the leader's side: inviting challenge so fewer people need to rehearse.

## About the makers

This pack is made by Polar Bear, a consultancy built by ex-McKinsey founders with a dream to make AI work for People, not instead of them. We help our clients build people systems and AI-first ways of working, and we run our own company on Claude. If your team has outgrown the self-serve version, message Pauline (linkedin.com/in/paulinebertry).