SABEX

Inspiration

Having spent significant time working on the ground in both traditional schools and specialized sports academies, we witnessed a recurring, silent problem: student dropout. A student doesn't just quit overnight; they leave a trail of behavioral breadcrumbs—missed classes, erratic schedules, and fading engagement. Standard attendance systems just log the empty seats, offering zero proactive value. We were inspired to build SABEX to transform cold data into an intelligent shield, leveraging AI to alert educators and administrators before the student decides to walk away.

What it does

SABEX is an AI-native SaaS platform designed for educational institutions and academies. It goes beyond standard attendance tracking by acting as a predictive retention engine. By continuously feeding attendance logs and behavioral data into the Gemini API, SABEX identifies at-risk students in real-time. The system doesn't just alert the administrator; it autonomously generates tailored, empathetic communication drafts, providing actionable socio-emotional support strategies to re-engage the student and their family immediately.

How we built it

We engineered SABEX with a modern, high-velocity tech stack. The core architecture is built on Next.js and TypeScript, utilizing Supabase (PostgreSQL) for real-time data synchronization and secure authentication. The application is deployed on Vercel for optimal edge performance. The AI brain is powered by Google's Gemini API, which processes complex attendance arrays to output structured predictive insights and localized communication templates.

Challenges we ran into

Our main technical challenge was structuring raw attendance data into a format that Gemini could analyze effectively. We had to carefully engineer the system prompts so the AI could accurately distinguish between a normal excused absence and a true pattern of chronic disengagement, ensuring the alerts were highly accurate and useful for administrators.

What's next for SABEX

Our immediate focus during the Build with Gemini XPRIZE window is commercial validation. We are deploying SABEX in active local academies to gather real-world retention metrics and process live subscription payments. The goal is to prove our AI-native retention model with real paying users and demonstrable impact before the global deadline.

Built With

• gemini-api
• next.js
• postgresql
• supabase
• tailwind
• typescript
• vercel