Skip to content
Menu
Back

BonoboAI

Full Stack Developer – AI Integration / BonoboAI

A multi-tenant quoting platform for service companies. Operators paste a customer inquiry, an LLM extracts structured data (location, installation type, logistics), calculates line-item pricing against a product catalog, and generates a branded proposal ready for PDF export.

React, TypeScript, Supabase, Serverless Functions, AI/LLM, Tailwind CSS

Overview

BonoboAI is a B2B SaaS that turns unstructured customer inquiries into professional, itemized proposals. Operators receive a free-text request, the platform’s LLM extracts key parameters (address, service type, floor access, logistics), calculates pricing against a configurable product catalog, and outputs a branded proposal with line items, totals, and a marketing narrative — ready for review, editing, and PDF export.

Key Features

  • LLM-Powered Proposal Generation — Extracts 8+ structured fields from free-text customer inquiries, calculates line-item pricing, and generates a narrative proposal with reasoning per line item
  • Operator Inbox — Split-panel interface for managing requests by status (pending, in review, approved, archived) with tabbed detail views for summary, pricing, and AI Q&A history
  • Nearest Store Matching — Geocodes the customer address, calculates driving distance to the nearest company location via OpenRouteService, and applies distance-based surcharges automatically
  • Multi-Organization Support — Each organization configures its own product catalog, pricing rules, Q&A flows, and proposal templates
  • PDF Export — Off-screen rendering of branded proposals for client delivery

Technical Highlights

  • Supabase Edge Functions running two LLM variants (Moonshotai Kimi, Arcee Trinity) with structured JSON output
  • OpenRouteService integration for geocoding + driving distance with ZTL-aware routing
  • Zustand state management for the split-panel inbox UI
  • SWR for data fetching with cache invalidation on status changes
  • Configurable per-organization Q&A flows that feed into the LLM extraction pipeline