In Short

Bots++ is an add-on to Ed Discussion. It enables instructors to integrate the LLM (Large Language Model) of their choice with Ed Discussion. Bots++ can be used in moderation mode, allowing instructors to review the responses of the LLM before making them available to students. Instructors can upload course materials to the LLM and write custom prompts to improve its accuracy.

Please note: Instructors have to choose which LLM to integrate with Bots++ and is also responsible for the cost of the LLM. Bots++ is disabled by default. Please reach out to the McGraw Canvas Team and we will enable it for you.

Teaching Purpose

Bot++ can be used to

Encourage timely feedback – Provide students with quicker responses to common or straightforward questions, especially outside office hours.
Support scalable Q&A – Help instructors manage high volumes of posts in large classes by generating draft responses for review.
Reinforce course concepts – Draw on uploaded course resources so that responses align closely with your learning objectives.
Reduce redundancy – Provide consistent answers to recurring questions without instructors having to retype similar explanations.
Free instructor time for higher-order teaching – Let the bot handle routine or fact-based queries so instructors can focus on deeper conceptual guidance.

Step-by-step

Enabling and Configuring Bots++

The McGraw team will enable the Bot++ at the course level based on an instructor's request. Individual instructors will need to set up their preferred LLM provider with an API key.

Access Ed Discussion Settings
- From your Canvas course, click Ed Discussion in the course navigation.
- Click the gear icon at the top.
- Select Bots++ from the left-hand menu. (Please reach out to canvas@princeton.edu if you don't see this option).
Set a Disclaimer
- Under Disclaimer, add a short message that appears before every bot-generated output.
- Example: "This response was generated by AI and may contain errors. Please verify the information."
Choose an LLM Provider
1. After clicking on New Provider, you'll select the LLM provider of your choice (e.g., OpenAI or Gemini).
2. Then enter the API key and complete the required fields.
3. Click Save.
Configure Response Settings
Under Responses, choose who the bot should automatically respond to:
- Nobody – Only respond when manually triggered.
- Students – Automatically respond to student posts.
- Students and Staff – Automatically respond to all posts.
Then choose the response type:
- Answer – Provide a full answer to a question.
- Comment – Provide additional input without marking it as an answer.
  - Check Post as Staff Only (highly recommended) - Allow staff to verify the bot's response before making it visible to students.
Add Course Context
- Upload Course Resources (e.g., lecture slides, outlines, assignments) to improve the accuracy of responses.
- About the Course – Add a description and learning objectives. This text is included in the bot’s prompt.
Advanced Settings
- Core Prompt – Customize the base prompt the bot uses. Leave blank to use the default.
- Ignored Categories – List any discussion categories the bot should ignore.
- Throttling – Enable Per Student Cap to limit how often a student can trigger the bot (set both hours and maximum generations).
- Data Privacy – Check Anonymous User to ensure the LLM does not receive the poster's name.

Using Bots++

If Bots++ is set to “Nobody” (manual mode):

Go to the relevant post.
Click the “...” menu and select Trigger LLM Bot.
Review the draft response:

If acceptable, click “...”, then select Convert to Answer.
If not, click Regenerate to request a new draft.

In automatic mode:

Responses will appear without manual triggering, based on your configured settings.

Helpful tip: Always review LLM-generated responses before posting them to students to ensure accuracy. You might also want to keep the disclaimer visible so students understand the source of the responses.

Additional Resources

Ed Discussion Bot++ AI LLM Overview