Role-play personas can now be enhanced with full-motion video avatars that lip-sync to speech, shift expression, make eye contact, and react to what's being said — making practice feel more like a video call and less like a voice over a picture. This article covers how it's turned on and what you and your learners will see.
Overview
When Moving Avatars is enabled, the AI persona in a role play appears as a real-time video avatar instead of a still image. The avatar lipsyncs to speech, shifts expression, makes eye contact, and reacts to what's being said — so practice feels like a video call, not a voice over a picture.
The persona also perceives the learner: it reads facial expression, gaze, and pacing, and adapts its behavior in response. Scoring, transcripts, and the rest of the role-play experience work exactly as they do today.
Turning on Moving Avatars
Moving Avatars can be enabled or disabled at the org, team, or assignment level. Your Customer Success Manager helps switch it on for your organization; from there, you control it per assignment.
Turning it on for an assignment:
- Create a new assignment, or open an existing one and choose Edit.
- Go to Final Details.
- Under Advanced Settings, turn on AI Video Avatar ("A lifelike video avatar is shown for the AI persona during role play").
- Save the assignment.
Any role play started from that assignment will now use the moving avatar.
What learners see
- Before the role play begins, the persona appears as the static image that was selected for them — the same image you see today.
- Once the learner enters the role-play screen, the persona appears as a moving video avatar, animating as it speaks.
How the persona reacts
The avatar doesn't just move — it responds to the learner. It reads facial expression, gaze, and pacing and adapts in character: if a rep breaks eye contact or rushes through a key moment, the persona reacts accordingly.
Camera is optional. Perception works best with the learner's camera on, but if it's off it falls back to audio-only signals like tone and pacing — so a webcam is never required.
Recording and playback
- The role-play recording includes the avatar, so reviewers see the persona's on-screen behavior on playback.
- Recordings with an avatar can take a little longer to finish processing after the session, because the avatar video uploads as well. Once processing completes, playback includes the avatar.
- Scoring and transcripts are unchanged. Moving Avatars only changes how the persona appears — it doesn't affect how a session is scored or transcribed.
Watch it in action
FAQ
Do I need to do anything to turn this on?
Moving Avatars can be controlled at the org, team, or assignment level. Your Customer Success Manager helps enable it for your organization; from there, admins turn it on per assignment under Final Details → Advanced Settings → AI Video Avatar.
Is a webcam required?
No. Moving Avatars works with the camera on or off. With the camera on, the persona also responds to the learner's expression and gaze; with it off, it adapts to audio signals like tone and pacing.
Does Moving Avatars change scoring or transcripts?
No. Scoring and transcripts work exactly as they do without avatars.
Why does playback take a little longer with an avatar?
The avatar video uploads along with the session, so recordings can take a bit more time to finish processing. Once they do, playback includes the avatar.
What does the learner see before the role play starts?
The persona's selected still image. The moving avatar appears once the learner enters the role-play screen.
Need help? Reach out to your Customer Success Manager.