Posted on October 2, 2026 by Samantha Becker and Jake Erney
Three UT San Antonio undergraduates designed, built, and launched a pilot AI agent for their university. Then they used the experience to land internships in New York, Chicago, and Costa Mesa.
Three students building AI campus support.
What happens when three UT San Antonio undergraduates are given an emerging technology, a real campus challenge, and room to figure it out?
Meet MEEP, short for My Educational Engagement Partner, an AI agent built to serve as a front door to UT San Antonio support. Type a question into MEEP and, within seconds, it starts pointing you somewhere useful.
It draws answers from a custom-built knowledge base complete with useful information across nearly every institutional area. Whether you are a student, staff member, or faculty member, MEEP can help.
But what truly makes MEEP special is who built it. Three UT San Antonio undergraduates designed, developed, tested and helped establish the guardrails for the agent as part of the Students AI Partners initiative within Academic Innovation, a division of Academic Affairs at UT San Antonio. There was no predetermined answer, giving the students room to explore an emerging technology against a real institutional need and make decisions as they went.
By the end of the term, MEEP was ready for a faculty demo. More importantly, the students who built it had something increasingly valuable to bring into their own job interviews: real evidence that they knew how to wrangle emerging technology responsibly, with the help of a team.
Learning by building
The project moved quickly. It began in late February, and by mid-March, the team had mapped out the most frequent questions Academic Innovation receives and started building MEEP’s knowledge base. By early April, the agent itself was configured and ready to handle queries.
Three UT San Antonio undergraduates were behind MEEP’s development:
Paul Gonzalez, a fourth-year Medical Humanities major and Student Ambassador for Academic Innovation, led MEEP’s overall architecture and coordinated team operations.
Isabella Bartolomeo, a fourth-year Cybersecurity major, focused on technical development and refinement.
Hugo Pickard, a second-year Electrical Engineering major, worked directly with the generative and agentic systems behind MEEP.
For the students, the hardest part was navigating uncharted territory.
Official Copilot Studio documentation ran out long before their questions did. When standard guides yielded no answers, the team began developing its own methodology for building, testing, and reviewing the agent — then teaching one another what they learned each week.
That process turned MEEP into something much richer than a software exercise. The students were effectively running small-scale research and development. With no roadmap, they identified problems, tested approaches, discovered technical constraints, and established practices that future student teams will inherit.
They also defined MEEP’s boundaries. The agent does not solicit personal data, offer health recommendations, or assist with coursework. The team established these guardrails by consulting department directors and faculty regarding sensitive topics, then evaluating what they themselves would want an AI agent to know about them as students. Staff provided counsel, but the students made the final calls.
From experiment to experience to workforce
The initial investment was modest: roughly $4,000 covering three Copilot Studio licenses, three paid student internships, and MacBook Neos for the job. In return, the university gained a working support tool and a scalable template for practical student learning.
MEEP also arrives at a time when colleges are being asked to do more with limited resources. In a recent EAB survey of 459 student success leaders, a majority identified budget constraints and staffing shortages as major barriers to student support, while noting that AI could handle routine interactions so staff can focus on complex, high-touch needs.
That is the space MEEP is designed to occupy: acting as a digital front desk so staff can concentrate their energy on more complex questions.
The project also aligns with student feedback at UTSA. In recent RowdyPulse surveys, students cited practical experience and clear usage guidelines as key priorities for AI literacy. Building MEEP gave its creators direct exposure to those concepts — and that kind of hands-on experience is becoming increasingly critical.
As traditional pathways into early-career roles may be tightening, practical experience is becoming crucial. The share of salaried job openings open to candidates without prior experience has dropped 73 percent over four years, according to an analysis reported by Forbes — falling from roughly one in 15 postings in 2022 to about one in 50 today.
The same technology reshaping the first rung of the career ladder is the one three Roadrunners used to climb it. They didn’t just learn about AI in theory. They built something a university actually put in front of people, then walked into interviews with an artifact they could open and demonstrate. And it worked.
Gonzalez accepted an internship in Costa Mesa, California; Bartolomeo landed in Chicago; and Pickard is working in New York.
Making room for student R&D
MEEP demonstrates the value of transitioning students from passive AI consumers to active, ethical builders. It offers an exciting model for applied learning that breaks out of traditional classroom confines.
The experience also sets a repeatable framework for future student-led R&D within Academic Innovation at UT San Antonio. The inaugural team leaves behind a functioning agent, a structured knowledge base, development practices, and lessons the next group can build on. It is a first iteration, built quickly by three students who have since moved on to internships. The more interesting question is what the next team will do with it.
Developed within UT San Antonio Academic Innovation with support from Melissa Vito, Marcela Ramirez, Joe Tobares, Claudia Arcolin, Haley Craig, and the Academic Innovation Technology Solutions team.
References
Student Support Needs Outpace College Capacity, Inside Higher Ed, September 3, 2026.
Can AI Help Colleges Reach Struggling Students Earlier?, Inside Higher Ed, September 11, 2026.
What The Changes In Entry-Level Jobs Mean For New College Graduates, Forbes, August 2, 2026.
— Samantha Becker and Jake Erney
