Roosevelt University is a private university in downtown Chicago, founded in 1945 on a commitment to access. When its enrollment and IT teams sat down with CollegeVine, they didn't start by asking where AI could fit in.
They started by mapping the friction students and staff were actually running into: tasks eating an hour of someone's day, and requests that stalled because no single office could finish them alone.
"Instead of focusing on AI, we focused on problems we wanted to solve using AI," says Neeraj Kumar, Roosevelt's VP of Information Technology and CIO.
The first agent ran at the front door
A prospective student calling admissions used to mean a note in Slate and a wait behind dozens of other students for the same counselor. Roosevelt and CollegeVine's first agent picked up that call instead, running the full outreach arc across email, phone, and text. One call ran 8 minutes; the student, guarded at first, told the agent things the enrollment team doubted he'd say to a person.
Financial aid saw a drop in seasonal hiring
The aid office used to bring on part-time staff every summer just to survive its own call volume. The agent Roosevelt and CollegeVine built went live in April, trained on Roosevelt's own aid knowledge, and by May the office trusted it enough to see a real drop in seasonal hiring. This past cycle, it handled 66% of incoming calls and resolved 2,000 of 3,030 replies without an escalation.
A Slate analyst, built overnight
For enrollment leadership, CollegeVine built an agent that answers directly against Roosevelt's Slate CRM data. Roosevelt described what it needed and had a working agent by the next morning; a couple weeks of testing later, it was fully live. The team now uses it regularly to surface insights and act on them right away.
One front door for every office
Students don't always know which office to ask. Fala, Roosevelt's virtual one stop, answers for financial aid, the registrar, housing, IT, and more, all in a single conversation. In its first 10 days it handled 1,125 conversations with students, with 45% of them occurring outside business hours
What's next: one record per student
The agents live today all answer from a shared knowledge base, so every student gets the same answer to the same question. The next step is tailoring that answer to each student's own record, which is why Roosevelt and CollegeVine are building an ontology layer across Banner, Slate, and the systems around them.
.png)
.png)
.png)
