Role
Founder, Product Lead (Vision + UX), Builder (Implementation with AI tooling)
Product Type
AI-driven conversational web app (anonymous storytelling + insight)
Status
Live MVP, iterating based on real usage and feedback
Designing an AI-mediated space where people can candidly share workplace experiences—built with intentional tone, boundaries, and selective memory.
The Problem
People in difficult work environments often need to talk through what happened and sanity-check what’s “normal,” but the available outlets are flawed: public platforms reward polish, HR channels prioritize risk, and friends lack context. The result is isolation, rumination, and poor decisions made without perspective.
Hypothesis
I believed a conversational interface could lower the barrier to telling the truth—if it was designed to listen first, ask better questions, and avoid turning into advice or escalation. The hypothesis: tone, pacing, and boundaries would matter more than “smart” answers.
Key design principle: The product listens first. The most valuable output is clarity, not advice.
Goals
- Create a low-friction space to “get it out” without identity exposure
- Encourage reflection through follow-up questions, not solutions
- Capture lightweight, structured signals about companies/people (selective memory)
- Keep the experience safe, bounded, and non-clinical
Non Functional Goals
- Therapy, coaching, or prescriptive guidance
- Employer mediation or HR escalation
- Social networking, commenting, or virality mechanics
- Long-term storage of sensitive personal details
My Role & Ownership
I owned the product vision, interaction principles, and scope decisions; designed the conversational UX; and collaborated closely with AI tooling (including ChatGPT) to accelerate iteration across copy, prompt behavior, edge cases, and implementation details. I made the tradeoffs around safety boundaries, what to remember, and how the experience should feel.
Solution Overview
- Anonymous chat experience designed for candid storytelling
- “Doris” tone calibrated to validate and draw out specifics rather than solve
- Selective memory approach: focus on companies/people when relevant
- Iterative prompt + UX refinement based on observed friction and user flow
- Clear constraints to avoid unsafe or overly directive behavior
Key Tradeoffs & Constraints
I intentionally traded breadth for trust. Rather than adding features that could increase engagement but compromise safety (social mechanics, sharing, public profiles), I prioritized tone, boundaries, and clarity. I also constrained memory to reduce risk and keep the product aligned to its core purpose.
Why this matters: This project demonstrates product judgment around safety, scope, and trust—core competencies for senior product leadership.
Outcomes and Future
- Live deployment with early organic usage
- Repeat sessions (signal of trust)
- Qualitative feedback validating the “listen first” tone
- A concrete roadmap emerging from real conversational patterns