HIGHER EDUCATION • AI PRACTICE • SHARED LEARNING

Brett Pollak

Building and learning from institutional AI in higher education

San Diego, California

I work with colleagues at UC San Diego on practical AI systems like TritonGPT. This site is a place to share what we're learning about implementation, governance, adoption, and the institutional questions that matter once AI moves beyond experimentation.

If you're working on similar questions in higher education, I'd be glad to compare notes and learn from your experience.

Brett Pollak
WRITING & CONVERSATIONS

Writing and Conversations

Articles, interviews, and case studies across 30+ publications on AI and institutional change in higher education

Browse Writing & Media
IN PRACTICE

What We're Learning in Practice

Making Knowledge Easier to Reach

At UC San Diego, TritonGPT helps students, faculty, and staff find institutional knowledge more quickly in the course of everyday work.

73K+ users across campus

Reducing Administrative Friction

In workflows like contract review, TritonGPT has reduced turnaround time by 91%, giving teams more room to focus on judgment and service rather than repetitive review.

91% time reduction

Building With Campus Partners

Work like this depends on collaboration across service, infrastructure, analytics, and web teams supporting a campus community of more than 73,000 people.

80+ colleagues across technology teams
TOPICS

Topics I Keep Coming Back To

AI in Higher Education

How institutions can move from experimentation to useful, trusted practice without losing sight of mission, governance, or people.

Institutional AI Implementation

What it takes to move from pilots to real use across a campus, including infrastructure choices, governance, support models, and adoption.

Governance and Institutional Context

Privacy, security, stewardship, and making institutional knowledge easier to access without flattening the nuance behind it.

Adoption and Organizational Change

How teams build trust, involve stakeholders, and learn what actually helps people in their daily work once a tool leaves the demo stage.