Use a non-generative AI chatbot with a pre-programmed, human-curated knowledge base to answer student questions with pre-approved responses¶

Strategy · All strategies
Evidence · no claims cited

Description¶

The strategy pairs a non-generative AI that algorithmically matches incoming student texts to a campus-specific knowledge base of pre-approved responses, with human administrators who answer unmatched questions and add them to the knowledge base over time. This frees staff from answering commonly asked questions repeatedly while keeping responses aligned with university communication standards. The article notes that whereas generative AI poses risks related to hallucinations and data safeguarding, this approach lets the university maintain tighter control over responses and protect personal data.

Design Implications¶

Context¶

Requirements¶

• A chatbot team to develop, expand, and monitor the knowledge base
• Human system administrators to triage flagged questions

Constraints¶

• The AI cannot answer questions outside the knowledge base; those are flagged for human response

Target Learners¶

• Undergraduate students communicating via text message

Target Learning Goals¶

• Timely, accurate answers to administrative and academic questions

Related Strategies¶

• Implement educational chatbots with human-centred design, explicit AI policies, training, curated knowledge bases and continuous monitoring
• Govern tutor knowledge through teacher-curated document corpora with RAG and document-driven adaptation
• Deploy an educational chatbot to give students instant question-and-answer support outside class when teacher interaction time is limited

Examples¶

-

Key Sources¶

• Mata, Catherine, Emily Russell, and Lindsay C. Page. (2026). Sustaining AI-Enabled Student Support: A Four-Year Implementation and Impact Study. (EdWorkingPaper: 26-1409). Annenberg Institute at Brown University. https://edworkingpapers.com/ai26-1409