01
Routine work, not human work.
AI can help explain terms, collect context, and compare routine options. Policy exceptions, distress, and consequential life decisions need an accountable person and a specific path to reach them.
For advisors
Built with advisors, not instead of them.
Academic advisors do some of the most important work on a college campus. They help students figure out whether to switch majors. They notice when a student stops responding and reach out before it becomes a crisis. They know their students.
Project Ardi tested whether a conversational planning tool could help students do more of the preparation before an advising meeting. The product is now closed, but the boundary it exposed still matters: software can organize routine complexity; it cannot replace institutional judgment or the relationship at the center of advising.
Three commitments
02
Visibility for advisors, not surveillance of students.
A future system should let the student choose what to carry into an advising meeting. The student's own words, source records, and unresolved questions matter more than a hidden machine profile.
03
No predictive risk scoring. Ever.
Ardvarq's position is to personalize requested help around student-stated goals and context, not to rank students by predicted failure. Support should expand the choice space, never quietly narrow it.
Go deeper