Work in progress

Conversation Guardrail — Conversation Intelligence Prototype

A personal conversation-intelligence concept for reading exchanges more accurately, recognizing reciprocity, and improving response timing without losing an authentic voice.

The problem

Conversation coaching can become generic, overconfident or overly strategic. Real exchanges need verified context, fair interpretations and advice that remains proportionate to what was actually said.

What I built

Developed the Conversation Intelligence V1.2 specification, a source-traceable reconstruction workflow and an editorial app-interface concept. The screens are visual prototypes; a functional standalone app has not yet been released.

What you can see now

Canonical FEATHER / Grindr incident dossier — 45 transcript entries reconstructed from a screen recording and eight screenshots, with gaps marked
Conversation Intelligence V1.2 — documented behaviour, pacing, reciprocity, provenance and interpretation guardrails
Editorial application mockup — exploratory screens and illustrative sample data, not live user metrics