Utah describes consent and advisor review for AI appointment notes; benefits remain unmeasured

Advising directors may face documentation burden and inconsistent follow-through. Include advisors, students, accessibility staff, records owners, IT and counsel in discovery. Ask how long note creation and correction currently take, which errors matter, and whether students can decline without losing service. A credible value hypothesis is better documentation with less administrative effort, subject to measurement. Offer a small consent-based pilot and review the complete workflow before licensing expansion. The university account supports a concrete process discussion but supplies no quantified return. Do not promise retention improvements, regulatory compliance, staff reductions or savings from generated summaries alone.

Map the path from appointment capture to draft summary, human correction and the official record. Prerequisites include approved meeting settings, accurate student identity, permitted data categories and records retention rules. Validate access boundaries and deletion for intermediate artifacts as well as final notes. Confirm whether transfer is manual or supported by an approved integration; do not assume the source establishes an API.

Proposed proof of value

use synthetic cases to test missing deadlines, incorrect advice, consent withdrawal and corrections before a limited live pilot. Measure end-to-end time including review. Autonomous record updates should require a separately justified design.

Campus Advising Solutions or its local equivalent should own the service, advisors should approve notes, and records and privacy teams should own data rules. Train staff in consent, correction and handling requests to stop capture; keep an equivalent manual route. Dependencies include settings verification, approved retention and a student complaint process.

Proposed acceptance criteria

every sampled final note has documented advisor review, all consent-withdrawal scenarios work, and error and total documentation-time measures meet locally agreed thresholds. These are proposed gates, not Utah results. Watch for review fatigue, selective adoption and uneven transcription quality before extending the workflow.