Why it is different
The AI disagrees
with you on purpose
A roleplay partner that accepts everything you say teaches nothing. These scenarios push back, go quiet, get defensive and change the subject, because that is what people do.
- The character holds a position and has to be moved, not just talked at
- Scored against named criteria, with evidence quoted from the transcript
- A score of the attempt, not a judgement of the person
- A manager review queue on Pro, so a human signs off before it counts
- A readiness report that says who could actually handle the conversation
- Weak criteria turn into course recommendations automatically
Before you ask
What people check first
Who can see somebody's practice transcript?
The learner, and their manager. It is not published, not on a leaderboard and not visible to their peers. The leaderboard ranks effort rather than scores, for exactly this reason.
Can somebody farm one easy scenario?
No. Repeat attempts at the same scenario are capped per month, which pushes people across the library instead of replaying a favourite.
Is voice practice included?
On Pro and above. Voice costs several times more to run than text, so charging it the same would be a quiet subsidy paid by everybody who does not use it.
Which plans have this at all?
Grow gets text practice and scored evaluations. Pro adds voice, manager review and the readiness report. Core has none of it.