Project Context
Writing Assistant keeps editing support inside the document. Users do not need to jump into a separate chat to improve text.
CASE STUDY
A selection-based document editing workflow where users highlight text and apply AI actions without leaving the editor.
Writing Assistant keeps editing support inside the document. Users do not need to jump into a separate chat to improve text.
I designed the interaction model and interface structure for the workflow, focusing on how users select content, trigger an action, and review the result without losing context.
The interaction rule is simple: selection equals intent. Highlight text, choose an action, review output.
Three compact states show how the same selected text can be transformed in different ways.
Converts selected text into another language while keeping the original document visible.
Condenses long paragraphs into shorter, easier-to-read content.
Transforms dense text into structured bullet points for faster scanning.
I designed the flow as a loop: select, act, review, continue. It feels intuitive now and leaves room for future actions.
The assistant panel works as a contextual editing layer. It does not replace the document editor; it supports it by making common AI transformations available at the moment they are needed.
The editor now supports writing decisions in real time, with AI actions placed exactly where users revise, compare, and continue drafting. This made the feature feel less like a side tool and more like a natural part of the writing process.