Campus AI workflow
PilotFaculty pull approved activity data into a BioBib draft, then check every section before it goes anywhere.
- 20Review partsSection-level routing
- .docxOutputBioBib draft
- UCSDWorkflowAcademic personnel
Accountability
How this use case is governed
- Service owner
- TritonAI solutions team and the sponsoring academic service owner
- Human oversight
- The faculty member or authorized reviewer validates every section before submission.
- Measurement plan
- Draft completion time, missing items, corrections, and reviewer satisfaction.
- Data boundary
- Public description; faculty records follow approved service controls
Product media
See the workflow in action
Silent screen recording showing a Word CV upload, section-level review, TritonAI routing, and a downloadable BioBib draft.
- Word CV Input
- Section Review
- TritonAI Routing
How it works
How the workflow fits
The problem it addresses, what it actually does, and how far along it is.
Problem
Putting a BioBib together means sorting, categorizing, and formatting the same kinds of record over and over, then checking all of it.
Solution
The pilot splits a Word CV into 20 parts and runs each one through TritonAI for formatting and extraction. You get back a .docx BioBib draft laid out in the expected UCSD biography and bibliography sections. It will point out items that look missing or ambiguous, but it does not claim the draft is complete or correct.
Current status
Pilot. The faculty member or service owner is still the source of truth for the final record. It all runs in the browser, so faculty and academic personnel teams go from upload to download without installing anything.