Design
Designing AI Interfaces Around Intent, Not Text Boxes
Scope AGI 2 min read
Published
Expanded
Most AI systems start with a blank text input. We discuss why understanding the user's high-level intent first—and then tailoring the visual workspace to match that objective—is the key to human-centered AI design.
An input is a beginning
A text field is a flexible way to express a need, but it does not have to define the shape of every result. A quick explanation, a document, and a comparison ask for different amounts of space and structure.
Designing around intent means paying attention to what a person is trying to accomplish after they press send. The interface should help them read, compare, inspect, or act on the response.
Choose structure that serves the content
Short replies should remain easy to scan. More substantial material can benefit from headings, cards, lists, and an associated file preview. These elements earn their place when they make the result easier to use.
Visual consistency matters across those formats. Typography, spacing, contrast, and a clear hierarchy let the content change without making each answer feel like a different product.
Preserve continuity between steps
A person may move from speaking to reading, from a result to its source, or from a conversation to a dedicated Space. Those transitions should preserve the relationship between the task and the information on screen.
This is the design direction behind Linse: begin with a clear expression of intent, then give the result a useful form. The later notes in this journal describe the concrete changes made toward that direction.