Project Workflow

How we review and support requests — moving from an early concept to an AI-powered solution for the Cornell community.

1
Submit an idea
Faculty and staff begin by filling out an intake form. The team then reviews the information to understand existing workflows, assess feasibility, and decide whether AI is a fitting tool.
2
Initial review
Each semester, the AI team gathers to evaluate submissions using a scoring matrix — weighing technical feasibility, data readiness, potential impact, stakeholder bandwidth, and ethical implications. Poor fits are redirected elsewhere.
3
Discovery call
The team meets with requestors to understand current processes, data access, subject matter experts, technical resources, and the broader goals and core problems to solve.
4
Selection and triage
The strongest proposals are chosen by feasibility and impact, limited by bandwidth. Complex, privacy-sensitive, or large-scale needs become formal projects, while others receive consultation or training. Remaining viable ideas are queued for the next cycle, and selected requestors are emailed.
5
Team assemble
The requestor gathers subject matter experts and end users. The AI team assigns a tech lead and Computer Science students by skills and interests, and OED may add a consultant for organizational design work. Weekly sprint meetings are set.
6
Project kick-off!
The full team — stakeholders, tech lead, consultant, and students — meets to explore where AI can help, pick a key problem, set a clear sprint goal, and build an actionable plan.

Have an idea for using AI in your work? Start here.