Most of my work lately has been building voice practice for learning teams.
It started with AI role-play, the best-known use of voice AI in learning and development (L&D), and quickly grew beyond it. With the same system I now build three kinds of module:
- Scenario practice: the AI role-play. A learner talks, out loud, with a simulated character in a situation they’ll face at work, and gets a written debrief against the scenario’s rubric after every run.
- Rep practice: short drills, like a thirty-second answer to a pricing objection, with feedback after each attempt.
- Reflection: the AI walks the learner through a concept, then asks them to think out loud about their own experience of it. This is modelled on Henry Mintzberg’s approach to management learning: “reflection on natural experience in the light of conceptual ideas.”
Put together, they make learning paths: reflect on a situation, drill the moves, then use them in a full scenario.
Most of what I build is custom. A client’s own scenes and rubrics can teach the AI to run the role-plays their facilitators already run, and scenarios can be built around a workshop so learners practise exactly what it teaches. I’ve built them for sales teams, leadership programs and financial advisors, in English, French and several other languages. The voice practice I build is sold through CoachingOurselves.
A few design choices I keep coming back to:
- Fixed or varied characters. When the relationship is the point, I specify the character in detail. When the situation is the point, I leave the character open and let the AI vary who shows up, so the seventh run doesn’t feel like the first.
- Results, not conversations. Managers see who has started, who is practising and who has met the bar, and which objectives people keep missing. They never see what was said, and reflection is never reported.
- What people say, not how they sound. The system doesn’t infer, label or score anyone’s emotions from their voice. Doing so at work or in education is prohibited in the EU.
Role-play is where most organizations start. The learning comes from what you put around it.