Integrating Chatbots in Course Design

Guiding Principle:  Always use Goals-Centered Design Principles to decide whether a bot fulfills your course goals and learning objectives, just as you would for other learning assets.

Tip: Use our DCE prompt generator tool (beta version) alongside this checklist (i.e. this document) to think through parameters for your bot configuration.

How do I copy a bot from one course to another?

There is no direct way to copy a chatbot between courses, but you are free to copy and paste the bot instructions and knowledge base from one bot instance to another.

What is "RAG" and how is it being used in the chatbot?

RAG stands for Retrieval-Augmented Generation—a method where the chatbot uses a special knowledge base to ground its responses. Without RAG, a bot only "knows" what it has learned from its original training. RAG allows you to privilege certain accurate information over the general noise of the internet, or to tell it details about your course and syllabus.

What is the intellectual property (IP) policy for material used for the system prompt and RAG?

Any documents uploaded for RAG in your chatbot must be either owned by the teaching team, used under an appropriate license, or are otherwise covered as fair use under the TEACH Act (i.e. the same policies that cover what you can upload to your Canvas site).

What data, according to the university security policy, can I use to configure a chatbot?

Only data classified as Level 1 or 2 (public or low-risk internal data) should be used to configure chatbots unless explicitly approved. Do not use any data that includes student grades, IDs, or personally identifiable information unless authorized. For Level 3+ data, notify the DCE Teaching & Learning team for guidance.

Can the chatbot adapt to students with different proficiency levels or learning needs?

Yes—many chatbot platforms can be designed to scaffold responses, offering varying levels of support based on how students interact with the tool. When well-configured, a chatbot can tailor explanations, provide prompts, or adjust complexity depending on the learner's background. You can instruct the chatbot to prompt the user to check their comfort/knowledge level. Consider including adaptive features in the bot's system prompt and testing with diverse users.

How much time and effort will be required to set up, train, and maintain the chatbot?

Setup time can vary widely. A basic configuration may take under an hour, while developing and refining a bot aligned with course learning goals can take several weeks and multiple human testers. Testing and iteration—especially for accuracy, tone, and instructional fit—requires time. The DCE Course Design team is available to support this process.

Is it okay to use chatbot outputs as part of student assessments or assignments?

Yes, but with clear expectations. If students are allowed to incorporate chatbot responses into assignments, they should comply with your policy on acceptable use, give proper attribution, and attend to privacy and other ethical concerns. Be mindful of potential plagiarism risks.

Can I customize a chatbot's tone or personality to match my course style?

Yes! Your instructions can define the chatbot's tone, voice, and personality through the system prompt or configuration settings. You can also upload samples of your writing style to the knowledge base. You can tailor it to be more formal, conversational, encouraging, or even role-specific (e.g., a mentor or lab assistant), depending on your course environment and learning goals. This is often one of the most effective ways to align the chatbot with your teaching style.

Who is my point of contact when there is a technical difficulty?

For technical issues, contact the team who supports the platform. Generally, tech support is available for faculty, but not students. The DCE Course Design team can assist with pedagogical and design-related questions.