Video Roleplay: Practice face to face with a lifelike AI avatar

Most sales and support conversations now happen on video, camera on, screen shared, reading a face for reaction as much as listening for words. Voice-only roleplay never touched that half of the job. Now it does.

‍

Why this matters

A rep who is comfortable on a phone call can still be thrown by a video meeting: where to look while sharing a screen, how to read a buyer's expression when they go quiet during a pricing slide, how to keep talking naturally while narrating a demo. None of that is trainable through audio alone, and setting up a human counterpart for every practice session doesn't scale across a team.

‍

How it works

‍

‍

‍

Launched in v3.2, video mode replaces the standard voice call with a lifelike AI video avatar. When creating or editing a roleplay agent, an admin sets the call type to video mode and chooses an avatar by name or appearance from the available set. The rest of the agent, roleplay type, scenario, and scorecard, is configured the same way as any other agent, with the option to add scoring criteria specific to video: gestures, demo delivery, and what's said while sharing a slide.

‍

One tradeoff worth knowing: Gatekeeper, the screener a rep normally has to get past before reaching a voice roleplay's buyer persona, isn't available in video mode, so a video session goes straight to the persona without that initial screening step.

‍

When a rep starts the session, the avatar appears as a talking head that responds in real time, in voice and video, to what the rep says and how they say it. During the call, the rep can open screen share to walk through slides or a live product demo and narrate while the avatar continues to react and ask questions. Feedback after the call is scored against whichever criteria the admin configured, including gestures and demo delivery where applicable.

‍

Setup details are in Create a Video Avatar (Video Mode) Roleplay Agent.

‍

What this unlocks

A sales team can now rehearse an actual product demo, screen share included, against a buyer persona that reacts to pacing and tone, not just a script read aloud. A customer success team preparing for a QBR on video can practice reading silence or hesitation on the other end, something no phone-based roleplay ever surfaced. In real estate, where a first client meeting is often a video walkthrough of listings, agents can rehearse the visual and conversational parts of that meeting together instead of separately.

‍

A university admissions counselor can rehearse a virtual campus tour info session the same way it actually happens, sharing slides of housing options and financial aid timelines while reading a prospective family's reactions over video, rather than treating the walkthrough and the conversation as two separate things to practice.

‍

The bigger picture

Half of what makes a video meeting different from a phone call is invisible to a voice-only roleplay by definition: no face to read, no screen to share, no moment where a demo slide changes what a buyer's expression is doing.

‍

None of that gap closes by getting better at phone calls. Video is where most of this work already happens. Training finally happens there too.

‍