Manager Practice · Holding People Accountable
Who Is Accountable
You are a manager, and you need to hold someone to a commitment when they are blaming the system.
Voice-to-voice AI roleplay for leadership development
Designing peer learning since 2007
Your people speak out loud to a character who answers back, then read a written play-by-play of how it went. Every scenario is built from your situations, in your vocabulary, against the standard you set.
Try one now
2 scenarios are open. You are the manager, the character speaks first, and there is nothing to prepare — a hello, a quick mic check, and you are in it.
Manager Practice · Active Listening
What They Haven't Said Yet
You are a manager, and someone on your team asks for two minutes about something small — the real reason they came is underneath it, and they have not said it.
What you get afterwards
A written play-by-play of every run
When the conversation ends, you read it back turn by turn, with a note on the moments that mattered and one thing to try on the next run.
Voice-to-voice, start to finish. You speak and the character answers out loud, the way the real conversation would go. Nothing is typed, and the written play-by-play only arrives once you have finished.
Play-by-play · Who Is Accountable
Excerpt
Sam
Thank you. So you see it’s not really on me.
You
I do see how hard they are to deal with. But the contract was yours, Sam. The emails were part of the work. The deadline was the outcome, and the outcome was yours.
This was a very clear and firm re-establishment of Sam’s ownership of the final outcome, separating effort from accountability. Excellent.
Sam
That feels a bit unfair. I can’t make Legal work faster.
You
You can’t. What I’m wondering is, when you saw it slipping, what stopped you from escalating before the deadline?
You didn’t take the monkey here, and instead pivoted directly to asking about escalation before the deadline, which is exactly what was needed.
To try next timeNext time, after Sam shares his plan with the vendor, you could ask him what he’ll do differently to prevent similar cross-functional delays from impacting his deadlines in the future. This would help him solidify his ownership of the process, not just the immediate fix.
Built with you
What teams ask us to build
- Make a programme transfer. People practise the conversation the course taught them, in the week they actually have to have it.
- Move an engagement item. The feedback a manager has been postponing, said out loud to someone who does not take it well.
- Land a change the managers did not write. Rehearsed the week before the cascade, rather than debriefed the week after.
- Get a new cohort credible quickly. The conversations a new manager has to survive, had once here first.
We build them from your material: the survey items, the deck, the notes from the last round of exits. Or from interviews with the people who do the job. We write them with your trainers, in your vocabulary. What counts as landing it is agreed before anybody practises.
More than roleplay
Three kinds of practice
A programme is usually built from more than roleplay. We choose the mix with you.
Scenario practice
A spoken role-play with someone who reacts the way a real person would. Rehearse the conversation before you have it for real.
The kind you can try abovePractice repetitions
One short move, out loud — an opening, an objection, a piece of feedback. You get it scored, with one thing that worked and one to tighten, and then you go again.
Reflection
Fifteen minutes on a situation of your own, with a named idea set beside it. The guide asks the questions, you do the thinking, and a written reflection appears at the end with one thing to try.
As a programme. Scenarios are assigned each quarter, chosen with you, and sequenced from easier to harder so each one prepares the next.
What a pilot looks like. One to two hours a month, for three months, to work through six to eight scenarios.
For the manager and the training team
See who has practised, and who has met the bar
The manager gets a readiness view: who has practised which scenario, how many attempts it took, and who has met the bar. The bar is one clean run, with every objective landed in a single attempt.
It is not a score and it is not a ranking. The manager never sees a transcript: voices are processed live and never saved.
Questions L&D teams ask
What is AI roleplay?
You speak out loud with a simulated character who answers back, voice to voice, in a situation you will face at work. Nothing is typed. When the conversation ends you get a written play-by-play: the conversation turn by turn, a note on the moments that mattered, and one thing to try on the next run.
How is a run scored?
Against the scenario’s objectives. The AI reads the transcript of the run and marks which objectives you landed and whether the conversation reached its end; it scores what was said, never how it sounded. Every scenario on this page has a pitfall, a make-or-break mistake, and a run that trips one counts for nothing. Someone has met the bar when one run lands what the scenario asks, reaches the end of the scene and trips no pitfall, and their best run counts. There is no leaderboard and no ranking.
What do managers see?
Results, not conversations: who has practised which scenario, how many attempts it took, who has met the bar, and which objectives are missed most. They never see what anyone said, and reflection is never scored.
Is it only roleplay?
No. There are three kinds of practice: scenario practice (the AI roleplay), practice repetitions (one short move, scored, with one thing that worked and one to tighten), and reflection on a situation of your own. They can be sequenced into a learning path, from easier to harder. Only scenario practice can be tried on this page.
How are the scenarios built?
From your material: the deck, the rubric, the notes from the last round. Or from interviews with the people who do the job. We write them with your team, in your vocabulary, and agree with you what counts as landing it before anybody practises.
What languages does it run in?
The scenarios on this page run in English, French, Brazilian Portuguese, Japanese and Slovenian, and a scenario can be built in any of 97 languages. The dashboards are in English and Quebec French.
What happens to my voice and what I say?
Your voice is processed live and never saved to our servers. A practice run keeps no recording and no transcript: the play-by-play is written for you and not stored. We build no voiceprint, and nothing about your emotions is labelled, scored, stored or reported. The privacy notice has the details.
How do I try it?
Pick a scenario above. You give your name and work email, and you are in the conversation: no account, nothing to install, just a microphone.
If it is a selling conversation. Ramp on somebody who has to be credible in three weeks. A discovery call that never reached the person who signs. A price held in the last week of a quarter. Same build, written in that vocabulary instead of this one. For sales enablement →
Contact us
Send us a message and we will get back to you. If you tell us what you are working on, we will come back with what we would build and roughly what it would cost.
Thank you — your enquiry has been received. We will get back to you shortly.
Or, if you would rather put time in the diary now, book a conversation with Warren.
Languages. The platform interface is in English and Quebec French. This demo and its practice scenarios also run in Japanese, Brazilian Portuguese and Slovenian — and a scenario can be built in any of 97 languages.
Based in Montreal.
More from CoachingOurselves The design of AI voice-powered learning · AI roleplay in learning and development