Experimentation

Innovation & Pilots

Explore active AI pilots, learn from completed experiments, and propose your own. UMS encourages a culture of experimentation within clear data-security guardrails.

Case Studies

UMA

Active

UMA: $400K Davis Foundation Grant for AI & VR

Two-year grant to expand secure enterprise AI tools and VR instruction, redesigning at least 20 undergraduate courses across nursing, aviation, business, education, computer information systems, and music technology.

2026–2028 $400,000

Outcomes & Metrics

20+ courses redesigned with AI/VR integration
New VR lab infrastructure
Faculty professional development program
Student learning outcomes assessment framework

AI Integration VR Instruction Course Redesign Grant-Funded

System-Wide

Completed

Scribehow Process Documentation Pilot

Deployment of Scribehow across administrative units for automated process documentation. Staff create step-by-step guides by simply performing the process while Scribehow records and generates documentation.

2024–2025 Licensed

Outcomes & Metrics

31 active users across multiple campuses
Hundreds of SOPs documented automatically
Significant reduction in documentation time
Improved onboarding for new staff

Process Documentation Administrative Efficiency SOP Creation

System-Wide

Active

ChatGPT Team Deployment

System-wide deployment of ChatGPT Team licenses, growing from initial pilot to 175 active users. The most widely adopted AI tool in the UMS portfolio, used for writing, research, code generation, and data analysis.

2024–Present Annual License

Outcomes & Metrics

175 active users (largest tool adoption)
Cross-campus deployment model established
Data security protocols validated
User satisfaction surveys conducted

Enterprise AI Writing Research Code Generation

Lessons Learned

Multi-Tool Strategy Over Single Vendor

Procurement

The initial enterprise GenAI RFP process revealed that no single vendor could meet all needs. UMS pivoted to a multi-tool portfolio approach, which better serves diverse use cases across academic and administrative functions.

Adoption is Broad but Shallow

Adoption

73% of users rely on a single tool. While 294 users across 7 tools shows healthy breadth, deeper engagement and multi-tool literacy remain goals. Training should focus on matching tools to specific workflows.

Faculty Autonomy Drives Engagement

Pedagogy

Pilots with the highest faculty engagement gave instructors full control over AI integration level. Mandated approaches saw resistance; opt-in models with strong support infrastructure performed best.

Data Classification is a Prerequisite

Security

Early pilots revealed gaps in data classification awareness. Before expanding AI tool access, campuses need clear guidance on what data can and cannot be processed by each tool tier.