Generating Campus Solutions, One AI Agent at a Time

Promotional graphic for The Tulane Impact Studio featuring a large group gathered on a staircase. Text identifies them as the “Inaugural Promptathon Cohort, November 2026.”

Tulane’s inaugural Impact Studio Promptathon brought campus employees together to turn everyday institutional challenges into early AI-enabled prototypes. Across six teams, participants explored how conversational agents and lightweight interfaces could help people find resources, navigate policies, reduce administrative work, and make better use of institutional knowledge. The event’s central lesson was practical: begin with a real user need, define a focused use case, test it against realistic situations, and keep people at the center of the design. 

The projects addressed a range of campus needs. IRBee Builder, the agent selected as the judges’ top project, was designed to help researchers organize IRB application responses, using research materials and policy guidance to address reviewer feedback more effectively. The IRBee team emphasized that AI-generated content still requires careful human review. The Wave Exchange agent, selected as the Runner-Up project, proposed a campus marketplace where departments could list and find surplus property and find items they need. The concept aimed to reduce storage costs and waste while making useful assets more accessible for departments with limited or restricted budgets. The Wave Exchange prototype included photo-assisted listings and gamified points. 

two group photos of promptathon participants posing in front of a colorful mural.
Runners-Up: Wave Exchange and People's Choice: WellieGator teams pictured with judges.

WellieGator, the People’s Choice project, envisioned a wellness navigator for students, faculty, staff, and alumni. The proposed tool paired an agent with a dashboard to surface relevant wellness resources, tailor options to an individual’s goals and location, and encourage engagement through small activities and points. The WellieGator team also emphasized a clear boundary: a chatbot can help people locate mental health resources, but people experiencing serious distress should be directed to human support. 

Other teams focused on employee development, policy navigation, and IT support. The Performance Review Self-Review Coach was designed to guide eligible employees through the annual review process by helping them organize accomplishments and connect them to competencies – not by writing or submitting evaluations for them. 

Collage of Tulane Impact Studio Promptathon participants: team portraits in front of a colorful mural, alongside photos of attendees presenting prototypes and collaborating in a workshop
Self-Review Coach team pictured with judges

The Compass project was designed to make university-wide administrative policies easier to navigate, generating checklists and identifying offices that could answer follow-up questions. The team saw potential for compliance and for supporting staff who manage policy review cycles, increasing awareness among individuals seeking clarity through step-by-step processes. The TulaneNow-Ripple project sought to guide users through IT or facilities issues using jargon-free language, help them identify the appropriate ServiceNow forms, and connect them to personalized support channels more quickly.

Several key lessons from this milestone event are especially relevant as Tulane’s academic, administrative, and research communities adapt to an evolving AI landscape. First, useful AI projects begin with a specific point of friction and a clearly defined audience. Teams improved their prototypes by testing realistic scenarios, revisiting assumptions along the way, such as eligibility requirements for the self-review coach. Second, source quality matters: agents need complete, readable, authoritative information. Some teams found that dynamic webpages were not reliably accessible and instead used uploaded documents, such as PDFs.  Third, a prototype is a starting point rather than a finished service. The Promptathon sprint quickly enabled teams to demonstrate proof of concept while also revealing the need for further testing, reliable integrations, and phased implementation. Finally, responsible design requires clear boundaries around privacy, sensitive information, and high-stakes decisions – while still preserving human review and referral pathways. 

Collage of Tulane Impact Studio Promptathon participants: team portraits in front of a colorful mural, alongside photos of attendees presenting prototypes and collaborating in a workshop
Compass team pictured with judges
Collage of Tulane Impact Studio Promptathon participants: team portraits in front of a colorful mural, alongside photos of attendees presenting prototypes and collaborating in a workshop
TulaneNow team pictured with judges

The Promptathon offered a model for cross-functional, hands-on AI learning – bringing together individuals who understand a problem, giving them time to build and test a small solution, and making space to discuss both its promise and its limitations. The event celebrated practical prototypes while underscoring that successful campus AI depends on trustworthy information, thoughtful and inclusive scope, iterative testing, and institutional collaboration supported by university leadership. 

Participants were selected from applicants across the institution and worked in teams guided by facilitators Dr. Blaine Fisher and Sam Centeno of the Innovative Learning Center, as well as Brandi Francis and Garrett Broady of CloudForce. Teams presented prototypes to their peers and were evaluated on impact, feasibility, prompt quality and guardrails, innovation, and the strength of the use-case story and potential pilot path. 

Special thanks to the judges: Patrick Norton, Tulane’s chief operating officer; Noel Wong, Tulane’s chief information officer; and Brian Dreyer, CloudForce’s NebulaONE campus partner.

 

Congratulations to Tulane's Impact Studio Promptathon Participants! 

 

Participants of the 2026 Cohort include: Derek Amaya, Blake Autin, Mary Barron, Meredith Beers, Reinhard Breckner, Victoria Breyer, Susan Cheng, Gregory Clouatre, Tim Coady, Robert Connor, Timothy Craft, Karen D'Arcy, Brian Deskin, Stacey Eharb, Nadia France, Dani Galloway, Kevin Grant, Evan Hart, Lindsay Hellwig, Tiesha Johnson, Denise Kidd, Monique LeCompte, Christi Longlois, Manny Ocasio, Hemanth Pasula, Adriana Quiroga Velasquez, Carmen Sarduy, Julie Smith Price, Jarrod Sullivan, Angela Sutton, and Brett Wakefield. 

 

 

TulaneAI promotional graphic featuring a cheerful astronaut mascot in a Tulane shirt, with a QR code and the words “Scan. Ask. Discover.”
For more information about the Tulane Impact Studio Promptathon or TulaneAI, send an email to tulaneai@tulane.edu.