Campus AI workflow
ProductionProcurement staff get contract language compared against approved UC legal positions, marked up and ready for a qualified reviewer.
- 120→11 minReview time91% time savings
- 50+Risk categoriesLiability, IP, data, SLAs
- 3Contract typesNDA, T&C, Software
Accountability
How this use case is governed
- Service owner
- TritonAI solutions team and the sponsoring legal service owner
- Human oversight
- A qualified reviewer approves findings and every proposed change before use.
- Measurement plan
- Review time, completeness, reviewer agreement, rework, and escalations.
- Data boundary
- Public description; contract data follows the approved production controls
Product media
See the workflow in action
Silent screen recording showing monitored intake, processing status, and delivery of a redlined document for human review.
- UC Legal Position
- Policy-backed Redlines
- Tracked Changes
How it works
How the workflow fits
The problem it addresses, what it actually does, and how far along it is.
Problem
Every contract has to be compared against approved positions, with issues spotted consistently and every proposed change recorded. The first pass is repetitive, but it still takes expert judgment, so the queue backs up.
Solution
The workflow pulls out the relevant clauses, compares them with the approved playbooks, and assembles findings a reviewer can act on. It supports the reviewer. It does not approve or sign anything.
Current status
In production for UC San Diego Procurement. The workflow runs through a monitored inbox and portal. It ingests contracts, extracts clauses, compares them against UC Legal Position and university template terms, and produces annotated redlines for human review. Measured outcome: 91% time savings on NDA/T&C review (120 min to 11 min average).