Single-purpose product framing
I kept the product tightly focused on reply generation instead of letting it sprawl into messaging analytics, inbox management, or multi-tool automation. The narrowness makes the value clearer.
AI-assisted communication tool
Message Reply
An AI-assisted reply tool designed to reduce repetitive rewriting while keeping the response usable and on-voice.
Role
Product thinking, Android UI, interaction flow
Scope
Workflow framing, Android interface, response generation flow
Context
Solo project
Message Reply
Inbox
Draft
Suggested reply
Editable
SendEditRetry
Overview
A faster response workflow without flattening personality.
Message Reply focuses on one narrow but common problem: rewriting similar replies again and again. The product is intentionally small, with the experience centered on speed, clarity, and editability.
I noticed how much communication friction comes from answering the same kinds of messages over and over. I wanted to explore a lighter AI product that solves one clear problem without trying to be everything.
Timeframe
Narrow-scope mobile product concept
Challenge
A lot of daily messaging is repetitive, but manual rewriting still takes attention and time every single time.
Product angle
This project shows how I think about AI as workflow support: useful, focused, and respectful of the user's own voice.
Goals
What I wanted to achieve
Make the value obvious in a few seconds
Keep the user in control of the final wording
Avoid the feeling of a bloated AI wrapper around a simple task
Constraints
The product had to stay lightweight enough to feel useful immediately
The suggestions needed to feel assistive rather than generic or over-produced
The interaction had to work on mobile without becoming input-heavy
Process
How the project took shape
01
Strip the product down to one flow
The product only works if it is faster than manual rewriting, so I reduced the experience to a simple sequence: input, generate, review, adjust, send.
02
Design for editable output
I did not want the suggestions to feel too polished or detached from the user's own language. The interaction is built around giving the user something usable, not something final.
03
Treat AI as a narrow interface layer
The product framing stays intentionally modest. Instead of building around AI spectacle, I framed it around the practical reduction of repetitive effort.
Decisions
Product and interface calls that shaped the result
Output should invite editing
The suggestions are positioned as drafts, not perfect responses. That preserves the user's voice and makes the product feel more trustworthy.
Fast mobile interaction over feature depth
Because the product lives on Android, speed and friction mattered more than adding secondary controls or configuration options.