Voice Agents give students a conversational AI practice partner for speaking, debating, interviewing, and professional simulations. You design the mode, scenario, difficulty, and grading criteria.

Practice Modes

Creating a Voice Agent

  1. Go to Dashboard → Voice Agents and click New Voice Agent.
  2. Select a mode and configure the scenario details (company, role, topic, etc.).
  3. Set the difficulty level: Realistic (standard) or Tough (more challenging).
  4. Define the grading rubric — specify criteria (e.g., clarity, depth, accuracy, communication).
  5. Click Save. Optionally test the agent yourself before publishing to students.

Reviewing Student Sessions

After students complete sessions, review their full transcripts, AI-generated verdicts, criterion-by-criterion scores, and written feedback from the Voice Agent detail page.


What Makes a Voice Agent Different from a Chat Agent

Voice Agents are optimized for real-time spoken conversation. They use a low-latency AI model specifically tuned for back-and-forth dialogue, not long document retrieval. This means they are best for conversational practice — pronunciation drilling, interview simulation, debate practice, oral exam preparation — rather than answering detailed factual questions about uploaded readings.

<aside> 📌 Voice Agents do not currently use your uploaded course materials. They respond from general knowledge and your system prompt instructions. For resource-grounded answers, use a Chat Agent.

</aside>