First time using Dialogue? Start with the Quick Start Guide instead. Getting set up before a session? Use the Pre-Session Checklist. This guide is the full reference.

Dialogue is a conversational AI platform. Depending on how your instructor sets it up, a Dialogue interaction can involve voice, text-based feedback, and/or assignment submission, sometimes chained together in one flow. Because it often processes live audio, your environment and setup can have a big impact on quality.

Ways You Might Use Dialogue

A Dialogue interaction isn't always just a voice conversation. Depending on your course, it can combine:

• Voice. Speaking with a Dialogue AI persona in real time.
• Text-based feedback. Written responses or evaluation you receive back from Dialogue.
• Assignment submission. Uploading or submitting written work as part of the interaction.

These can be chained together into a single flow. For example, a course might have you submit a paper, then defend it out loud to a Dialogue AI persona, then receive written feedback based on that conversation. Your instructor will let you know what your specific course's Dialogue setup involves.

Environment & Setup

Treat a Dialogue interaction like a live meeting or oral assignment. The quality of your experience depends heavily on your physical environment and how you engage with the agent.

1. Use headphones whenever possible. This is the single most important step. Headphones prevent the agent from picking up its own audio output through your microphone, which is a common cause of the agent appearing to "talk to itself" or respond without being prompted. Over-ear or in-ear headphones both work; even basic earbuds make a significant difference.
2. Choose a quiet space. Avoid buses, coffee shops, or other noisy environments. Background noise, including a TV, radio, or nearby conversations, can be picked up by the microphone and mistakenly interpreted as your voice, causing the agent to respond at unexpected times.
3. Speak in clear turns. Ask a question or make a statement, then wait for Dialogue to finish its response before speaking again. Talking over the agent, or having loud background sounds while it's speaking, can cause mis-heard questions or trigger extra unintended turns.

Browser Recommendation

For the best experience, use Google Chrome. Dialogue relies on real-time audio streams that are most stable in Chrome. Other browsers may work but can introduce audio handling issues that worsen agent responsiveness or cause unexpected behavior:

• Safari: known microphone handling differences that can increase misfires
• Firefox: generally functional but less tested for real-time voice
• Edge: Chromium-based, usually reliable, but Chrome is preferred

If Something Goes Wrong

Common symptoms and what to try:

• Dialogue doesn't load, stalls, or starts "talking to itself"
  • Refresh the page
  • Move to a quieter space or mute nearby audio sources (TV, music, etc.)
  • Plug in headphones and start the session again
  • Switch to Chrome if you are on a different browser
• Dialogue AI persona responds when you haven't spoken. This is usually triggered by background noise crossing the agent's audio detection threshold. Headphones and a quiet space resolve this in most cases.
• Unexpected language switches or voice changes mid-conversation. These are known intermittent issues with the underlying voice AI platform. Refreshing and restarting the session is the current workaround.
• If the problem persists, follow your course instructions for reporting Dialogue issues (usually via your instructor or a support form).

What to Know About Visibility and Evaluation

Purpose: When you do a Dialogue activity, a voice or text conversation with an AI persona, that conversation is the activity itself. This section explains what's captured from it, why, and who can see it.

What's captured. The conversation itself (a transcript of what you said or wrote), basic session info (when you did it and how long it took), and how you engaged (for example, text vs. voice, or repeat attempts). If your course has attached a rubric, a score is generated based on defined criteria; that score goes to your instructor.

What's not captured. Your voice itself isn't stored. Conversations are transcribed to text; the audio recording isn't kept.

What you get after a session. When you submit, you'll see a short feedback screen: you can say whether the session went well or needs work, and optionally add a note on what you'd do differently next time. You can also download your own transcript from this screen. If your course allows it, you can also try the session again before it's treated as final; ask your instructor if you're not sure whether that applies to your assignment.

Why it's captured. This data exists to help your instructor understand what's actually happening in the activity, not to catch you out. A transcript and rubric score show your instructor the quality of your reasoning, not just whether you showed up. It's meant to support your instructor's judgment, not replace it; a score is a starting point for their review, not a final verdict on your work.

Who can see it. Your instructor, and others at your institution involved in the course (like an admin), can see this. It isn't shared outside your institution.

Does this replace grading? No. Rubric scores and transcripts inform your instructor's evaluation; they don't decide your grade for you. Check with your instructor on how much weight this carries in your course.

Dialogue setups vary by course. What it does in one class may look different in another, since instructors and institutions configure it for their own use.

Questions or Issues?

Follow your course's instructions for reporting Dialogue issues, usually via your instructor or a support form. For questions about how Dialogue is set up in your specific course (what it's used for, how it's graded, what's visible to whom), ask your instructor.