Digital Twins to Deliver Personalized Feedback
Presenter: Carolyn Brown, Instructional designer at Champlain College and faculty at Foothill College
Host: Casandra Silva Sibilin
About the presenter:
Carolyn Brown is a Learning Experience Designer at Champlain College, where she builds AI-powered assets for asynchronous courses. Her work spans embedded chatbots, feedback generation, and custom interactives used across disciplines.
Her recent project, Chauncey, is an AI advising assistant built on the Pria/Praxis platform and embedded in Champlain’s Student HUB. Chauncey reviews student milestone submissions, updates the gradebook, and generates personalized feedback in the advisor's own voice.
Carolyn is also an online professor at Foothill College. Teaching asynchronously shapes how she designs AI tools: she builds for the moments that support a student, and pays close attention to where it shouldn't replace a human relationship.
She works across a stack that includes the Anthropic API, Claude, NotebookLM, ChatGPT, and Gemini, and she's presenting on her Chauncey implementation at ICDLE 2026.
About the session:
As online higher ed continues to scale, advising teams face a growing capacity challenge: too many students, too few advisors, and too little time for the kind of personalized engagement that supports student persistence and success. This is a case study where an AI assistant was deployed within the LMS to extend the reach of a small advising team across a student population of several thousand.
While AI-powered chatbots providing 24/7 support are common in higher education, what distinguishes this implementation is the expansion of the assistant’s role beyond answering FAQs. Working in close collaboration with the advising team, Chauncey was designed to review individual student submissions for multiple assessments, which guide students from initial orientation through career readiness activities. Chauncey reads each student’s work, evaluates it against a rubric or checklist of expected elements, marks the assessment complete in the LMS gradebook, and generates fully personalized feedback.
Critically, this feedback is not templated: it is informed by the content of the student’s own submission, the assignment expectations, and a carefully crafted system prompt that encodes the tone and voice of the human advisors. That voice was developed collaboratively; advisors described how they communicate with students in face-to-face settings, and those qualities were translated into the AI persona. The result is feedback that reflects genuine institutional warmth, commending students for specific accomplishments and encouraging their continued engagement.
---
About the series:
Custom Bots/GPTs is a subgroup of AI in Education, a global group for sharing resources and information related to the impact of artificial intelligence on teaching and learning. Each month the subgroup meets to share strategies, network, and showcase educators doing innovative work with custom bots, GPTs, and apps. For more information and resources, visit edushare.ing/bots.