Inspiration Coordinators at Dress for Success were manually matching clients to volunteers over email and phone, we wanted to replace that with a system that gets the right person in the room, faster.
What it does A full-stack booking and rostering system where clients self-book appointments and an AI-powered two-layer algorithm instantly matches, ranks, and notifies the best-fit staff member, automatically cascading to the next candidate if they decline.
How we built it React + Material UI frontend, Node/Express backend with a JSON file database, JWT auth, and an OpenAI gpt-4o-mini semantic scoring layer on top of tag-overlap matching.
Challenges we ran into Naively calling OpenAI for all 32 staff simultaneously burned our quota in seconds, forcing us to redesign to a two-stage approach: tag-rank everyone first, then only call AI for the top 3.
Accomplishments that we're proud of The AI match toast which opens any booking and the system surfaces the top candidate with an AI driven reason and a one-click "Send Request" button.
What we learned Driving AI agentic programming in a team setting rather than alone.
What's next for Dress For Success - Staff Assignment System Real SMS/email notifications, workload-balanced scoring so assignments spread fairly across the team, and an AI intake step that auto-tags bookings from the client's own words before they hit the queue.
Built With
- next.js
- react
Log in or sign up for Devpost to join the conversation.