Forming intent
The user knows roughly what they want and has to turn it into something the system will act on correctly.
A blank field is maximally flexible and maximally unhelpful; a form is reliable and cannot express the unusual request. Most products need both paths in the same surface, with the structured one first.
9 patterns for this moment
Each one is a different way to resolve the tradeoff above.
A structured form that produces a good prompt, trading expressive freedom for reliable output.
A sentence with editable slots, so the grammar of a good request is visible and the user just fills blanks.
One input that accepts text, files, images, audio and screen captures as equal citizens.
Typed @-mentions that pin exact objects — files, people, tickets, tables — into the prompt as structured references.
Select content, get AI verbs scoped to that selection — rewrite, explain, extend, translate.
Greyed inline prediction the user accepts with a key, keeping the model subordinate to the typist.
The system rewrites a thin prompt into a fuller one and shows the rewrite before running it.
A surprise-me action that produces a valid result with no input, lowering the cost of the first try.
Post-response chips that propose the next reasonable move, keeping momentum without new typing.
Audit checklist for this moment
Walk one flow through these questions. Ticks are saved in this browser.