Students did not always ask clean, single-topic questions. Many messages combined admissions, financial aid, transfer credits, and next steps in one request.
Questions like this were difficult because the student was really asking about two different processes: application materials and financial aid.
The bot did not need to solve every account-specific issue. It needed to:
I focused on making responses clear, useful, and safe for a bot experience.
The response pattern followed a simple structure:
This kept ADA from overpromising while still helping students move forward.
This approach made bot responses more useful for real student questions, especially when users did not know which department or process their issue belonged to.
Instead of relying only on “I don’t understand” moments, the bot could guide students toward a self-service action or a human support path.
Students do not always use university language. They ask questions based on what they are trying to do next.
Designing for ADA meant turning internal processes into plain-language answers that helped students understand where to go, what to check, and when to contact a person.