BlackPrint Technologies · Aug 2026
BrokerAI
Conversational AI & Lead Qualification Engine
Overview
Architected and built a production WhatsApp sales agent and real-time operator dashboard that qualifies leads and books advisor appointments for two real-estate developments sharing one WhatsApp number.
Highlights
- 01
Built an OpenAI tool-calling orchestrator with a three-layer prompt (invariant voice core, per-project profile, turn context) and a deterministic reply validator that blocks dates and times the calendar never offered.
- 02
Engineered thread-level routing (click-to-WhatsApp referral, first-message fingerprint, history) with per-conversation lease locks, inbound message coalescing, Whisper audio transcription and atomic per-advisor booking through expiring slot offers.
- 03
Shipped a mobile-first operator PWA with live monitoring and human takeover, Telegram-first internal alerts with masked phone numbers, and a quality net of 13 golden conversation evals plus 475 Vitest cases across 43 test files.
Stack
- Next.js 15
- React 19
- TypeScript
- Turborepo
- Prisma ORM
- PostgreSQL (Cloud SQL)
- OpenAI API
- Whisper
- WhatsApp Cloud API
- Google Calendar API
- Server-Sent Events
- Vitest
Working on something similar?
Tell me about the problem and the data behind it. I reply within 24–48 hours.
Discuss a project