The challenge
Case managers checked in on participants by hand, texting and calling one by one. It took about 20% of their working time, and with busy caseloads, check-ins slipped exactly when they mattered most: right after a job placement.
What I built
An automated check-in system connected to Salesforce:
- Weekly text check-ins to about 200 participants, plus extra check-ins timed to job placement milestones
- AI reads every reply and sorts it: doing well, needs help, lost job, urgent request
- Replies that need attention alert the right case manager right away. Routine replies are logged without anyone lifting a finger.
- Consent is collected at enrollment, and every message and reply is recorded on the participant's record
Results
- Manual check-ins eliminated, returning about 20% of case managers' working time
- Over 80% of participants respond
- Problems like a lost job or a request for help reach the case manager the same day, instead of at the next scheduled call
- Every check-in is documented automatically