Haven

<aside> 🪴

AI-native Relationship Product · 0→1 Product Case Study

Tzu-Yang Tseng (Alan) · Founder & AI Product Builder

A recruiter-facing portfolio for understanding the product, my role, and how I work across product discovery, AI-assisted development, technical collaboration, QA, privacy, and iteration.

</aside>

In one sentence: Haven is an AI-native relationship product designed to help people understand important relational moments more clearly, choose safer and more useful next steps, and turn AI support back into real human connection rather than AI dependence.

What I’m building

Haven is an AI-native relationship product in NO-GO / outcome-open / pre-A01 release-readiness work. The product surface is coherent enough to demonstrate the current product architecture, but broader real-user exposure has not begun; bounded Founder QA / release closure is still open. Its current Minimum Lovable Haven is organized around six thin-but-real experience families:

• Understand — capture important moments, clarify what happened, separate observation from interpretation, and choose a human-owned next step.
• Re-Know — build a living, revisable understanding of a partner and relationship rather than a static quiz or deterministic score.
• Connect — support appreciation, contextual prompts, reciprocal discovery, and small actions that lead back to real interaction.
• Experience — help turn wishes and constraints into shared activities that people can choose, modify, decline, and actually do together.
• Repair — support safer pauses, conversation preparation, editable language, and post-conversation reflection without diagnosis or auto-send.
• Build — help partners explore shared hopes, rituals, meaning, and future direction while preserving disagreement and individual agency.

<aside> 🧭

Product principle: AI-native does not mean “AI everywhere.” The operating loop is context → evidence / uncertainty → smallest useful relational move → human choice / action → partner or reality response → outcome review → confirmed learning.

</aside>

My role

I am building Haven from 0→1 as Founder & AI Product Builder. My work spans:

• user problem definition and product discovery;
• product positioning and MVP / Minimum Lovable Product scope;
• feature prioritization and roadmap decisions;
• User Flows and information architecture;
• translating product intent into development tasks, Acceptance Criteria, and Test Cases;
• directing AI Coding Agents and reviewing implementation output;