AI AUTOMATION ENGINEER APPLIED AI
FORM PR-02 · PROJECT PR-0001 · Applied AI
A secure WhatsApp assistant that lets parents retrieve school information and authorized student records through natural-language conversations. It demonstrates agent orchestration, LLM routing, grounded retrieval, identity verification, guardrails, observability, and business-process automation.
Private data needs hard boundaries
Parents repeatedly contact school staff for routine information. Automation can reduce that workload, but student-specific requests introduce privacy and authorization risks that a general chatbot cannot safely handle without verified identity, relationship checks, and grounded data access.
Authenticated, intent-aware routing
| Step | System behavior |
|---|---|
| 01 | Receive the WhatsApp message through Twilio and match the sender to an authorized parent record. |
| 02 | Unknown or unverified users complete a six-digit OTP flow with a five-minute expiration. |
| 03 | Classify intent; route balances and attendance to PostgreSQL, and policy questions to the RAG pipeline. |
| 04 | Validate required fields and the parent-to-student relationship before returning protected information. |
| 05 | Return a concise answer, log the outcome, and send consequential write actions for human review. |
From message to verified answer
Technology · Intents · Safety controls