A public university we worked with supports its online student population through a Student Engagement Team of just ten advisors. Before deploying Risely, those advisors spent most of their time reacting to crises - students who had already decided to stop out, financial aid holds that had already delayed registration, engagement drops that had already turned into absences.
The team knew what proactive support looked like in theory. They just didn't have the capacity to deliver it at scale.
The Challenge
The institution's online student population skews heavily toward working adults - students with jobs, families, and irregular schedules who are more vulnerable to life disruptions than traditional 18–22-year-old students. These students benefit most from early, personalized outreach. They also receive it least, because they don't live on campus and rarely drop by office hours.
Prior to Risely, the team's proactive outreach capacity was approximately 150 personalized touchpoints per week across all ten advisors. That's one outreach message per advisor per three hours - including all the time spent researching student context, drafting messages, and logging notes in the CRM.
The Deployment
Risely deployed its AI Student Engagement Agent over a two-week integration sprint. The agent connected to the institution's SIS, LMS, and financial aid system, establishing a real-time behavioral monitoring layer across the full student population.
In the first week, Risely and the team collaboratively defined four intervention playbooks: first-week engagement for new students, mid-term check-ins for students showing LMS disengagement, financial aid deadline nudges for students with incomplete aid files, and re-enrollment outreach for stop-outs from the prior two semesters.
The Results
Within eight weeks of go-live, the AI agent was initiating 4,800 proactive touchpoints per week - a 32x increase over the previous baseline of 150. Student response rates to agent-initiated messages averaged 41%, compared to 9% for the institution's prior bulk email campaigns.
The advisor team reported that the nature of their work shifted significantly. Instead of spending time on routine follow-up, they focused almost exclusively on students the AI had flagged as needing human-level support: mental health concerns, academic appeals, financial hardship cases. Advisors described the change as 'getting to do the job we trained for.'
Preliminary retention data from the first full semester showed a 4.2 percentage point improvement in mid-term persistence compared to the prior year cohort - the largest single-semester improvement the institution had measured in five years.
What Made It Work
The institution's Engagement Team lead attributed the deployment's success to two factors: the quality of the agent's personalization and the team's willingness to iterate. 'The agent doesn't send generic messages,' she said. 'It knows which courses a student is enrolled in, whether they've logged into the LMS this week, and whether they have an outstanding financial aid task. That specificity is why students respond.'
The team reviewed agent performance weekly in the first month, adjusting playbook triggers and message templates based on response data. 'It's not a set-it-and-forget-it tool,' she said. 'It's a team member that gets better the more you invest in it.'