A first difficult conversation should not be the first attempt at one.
Chaplaincy training is generous with reflection after the fact and thin on rehearsal before it. ChaplainSim is the rehearsal.
Why it exists
You read about presence, you discuss a verbatim in supervision, and then you walk into a room where someone is angry, or frightened, or asking whether their mother is going to die — and you find out in real time what you do.
ChaplainSim gives you somewhere to find that out first, where the cost of a clumsy response is a fictional conversation going somewhere you did not intend, rather than a real person feeling unmet.
How the content is made
Written by a person, approved by a named chaplain, fixed by a checksum.
Every scenario is written by a person, then reviewed and approved by a named chaplain before it can be published. The approval is recorded against a checksum of the exact content, so a published scenario cannot be edited without going through review again.
AI was used to help draft material before that review. Every line was then edited by the named author and approved by the named reviewer. We say this plainly because a tool that teaches honesty in difficult conversations should not be coy about its own.
The reviewer is not independent. For this pilot, the chaplain who reviewed and approved the scenarios is connected to the people who operate ChaplainSim rather than being an outside auditor. The review is a real one by a qualified chaplain, and we would rather tell you that than let you assume otherwise.
Where AI is not used
No AI runs while you are using ChaplainSim. The patient’s words, the responses you choose from and the feedback you receive were all written in advance. No model is called when you start a scenario, take a turn, read feedback, or save a reflection.
Your reflections, your email address and your individual records are never sent to an AI provider and never used to train a model.
What we do not build
Absence is a design decision too.
No scoring
No grade, no percentage, no ranking against other trainees. Outcomes describe where a fictional conversation went, not how good you are.
No surveillance
A supervisor sees an attempt only when a trainee chooses to share that specific attempt with them.
No inference about belief
We do not profile you and would not know how to.