Chronus
Bringing AI roleplay to real-world conversation practice
Project Overview
Aubergine partnered with Chronus, a US-based enterprise coaching, mentoring and ERG platform, to bring AI Roleplay into its product. We designed and shipped a practice experience where employees rehearse difficult workplace conversations with AI personas and get structured, actionable feedback. It's built into the mentoring experience members already use, so they can practice without leaving the platform.
Impact
A clearer, more consistent in-vehicle experience designed to support faster and safer response operations.
3
modes text, live voice, and role-reversed practice
100%
of readiness bands reproducible from stored scores
70%
faster scenario authoring with AI-assisted setup
The Problem
The conversations that matter most get zero practice reps
.webp)
Chronus runs mentoring and ERG programs for global enterprises. Those programs help people prepare for the conversations that matter most, but they had no way to let anyone practise one. The first attempt usually happened live, unprepared, with a real working relationship at stake. Employees needed somewhere safe to get it wrong first. But adding safe practice took more than putting a chatbot in front of a scenario. An AI partner that drops character, flatters the learner or hands back a score nobody can explain does not build confidence. It teaches the wrong habits and puts the program's credibility at risk.
Personas that break character
Learners push hard, and a persona that slips teaches the wrong lesson.
Scores no one could defend
A grade a learner cannot challenge is a grade an admin cannot stand behind.
The Solution
AI roleplay built into the mentoring experience, making practice simple and accessible
Goals
Keep personas realistic and in character
Make every score clear and fair
Personas that stay in character and feel real
We built Roleplay into the mentoring platform members already use, so practice feels like part of their development rather than another tool to learn and its powered by Anthropic's models. Each persona is designed to feel like a real person, with its own goals and point of view. It responds to what the learner actually says and stays in character from start to finish, so practice feels close to the real thing.
Scoring learners can understand and trust
Feedback only helps if people trust it. Every result is grounded in the skills and objectives the organisation cares about, and learners can see what they did well and where to improve. Results are consistent and explainable, so learners understand why they got the result they did and admins can stand behind it.
Results
Turning difficult conversations from something people survive into something they can rehearse.
100%
of results traceable to the scores behind them
24/7
practice, with no partner or scheduling needed
70%
faster scenario authoring with AI-assisted setup
60+
ready-to-use scenarios in the scenario library
- Realistic practice with AI personas that stay in character
- Transparent scoring admins can stand behind and learners can question
- A ready-made scenario library, plus AI-assisted tools for creating new scenarios
No items found.
