Overview
Loom is designed for moments when someone needs to understand a potential partner, interpret a conversation, reflect on a date, or prepare the next interaction. It aims to support judgment without deciding whom the user should love.
COMPLETED
Loom: Your AI Relationship Coach
An AI coach that remembers relationship context and turns uncertainty into grounded, actionable communication.
2026
Project Snapshot
- Date
- 2026
- Role
- Product Manager, Front-end Developer
- Status
- Completed
- Launch
- Completed demo
My Role
As product manager and front-end developer, I defined the core journeys and translated the agent's memory, knowledge, and routing capabilities into a coherent interface.
Problem
Generic conversational AI can respond immediately but often lacks stable relationship memory, professional grounding, and a durable understanding of the user and each person they discuss.
Solution
Loom supports relationship profile analysis, screenshot-aware chat analysis, and structured date reflection. A curated relationship knowledge base grounds the advice, while separate short-, medium-, and long-term memories preserve the right context over time.
Core Flow
Receive a question, screenshot, or date reflection → validate relationship intent → route to coaching, chat analysis, date review, profile, or image mode → assemble profiles, recent turns, summaries, long-term memory, and RAG evidence → call at most one specialist Skill → stream actionable guidance → asynchronously summarize and update separate user and partner memories.
Outcome
A working demo now covers profile analysis, chat analysis, date reflection, and persistent relationship memory.
Media & Resources
