01Works

Three products, zero to market

Each case is told the same way I worked on it: half product manager, half designer — with the AI parts explained in plain words. Earlier, pre-AI work lives under Archive in the nav.

Beever Atlas

Enterprise AI knowledge · 2026 — present

An open-source memory system that turns scattered team chat into a permanent, cited, permission-aware wiki.

  • 0→1 distribution across six MCP registries and >2 million leads size
  • v2 direction: cross-source memory spaces with RBAC
  • Go to market and narratives for multiple layers
AI MemoryGraph memorySemantic memoryMCPRAGRBACGo-to-Market
Read the case
Beever Atlas Q&A chat interface
V-Note Live dashboard

V-Note Suite

Cantonese meeting intelligence · 2025 — present

Transcription, speaker labels, and template-driven minutes — one edition inside the customer's firewall, one whose bot joins your cloud calls.

  • 0→1 launch with paying enterprise & government customers
  • Three major B2B channels: Alibaba, HKT, HK government sectors
  • On-premises enterprise edition & Cloud edition
ASRLLM SummariesOn-Prem & On CloudMCPPricing
Read the case
Xprop landing page

Xprop (大廈通)

PropTech / property management · 2024 — 2025

A Cantonese-speaking WhatsApp chatbot for tenants and front-line staff, backed by a portal with dashboards, ticketing, and AI document processing.

  • Touch-based discovery workshops with JLL
  • >30% cheaper than incumbents · 50,000-building market
  • Designed the whole shipping journey
PropTechWhatsApp BotManagement PlatformGTM
Read the case