Use case
A product design review grounded in the interface you actually have
Try That UI puts the current experience, the proposed change, and the decision in one review. Product, design, and engineering are looking at that object, not at three retellings of it.
The failure mode this is for
A design review goes vague when the baseline disappears. Someone describes the current screen from memory, someone else pastes a proposal that was made from a different state, and the meeting debates a difference nobody can point at. The fix is mechanical: keep Current visible, put one proposed change next to it, and write down what was decided.
That is the whole product loop. Paste the interface, screenshot, or source. Scope the change so the two states stay comparable. Try an isolated component prototype on Proposed if the interaction is what you are judging. Then keep, reject, or share the decision. Captured screens do not replay the application, and the prototype is not the production code.
For product managers
Use the comparison to connect the proposed interface to the customer problem you are actually deciding. The question in the review is whether this change deserves implementation, not whether a new composition is attractive in the abstract. The decision record is what you carry out of the review: scope, what the preview showed, and the next step. It is not a metric dashboard, and it does not claim a result you have not measured.
For designers and design engineers
Make the visual and interaction difference inspectable. Current stays the screen you supplied. Proposed holds the change you want judged, including a component prototype when a still image cannot show the control. Hand off that pair instead of a verbal description of what moved. If the action you are proposing needs a recovery path, review it against the reversible actions note before you call it undoable.
For developers
Start from an approved direction with the screen and the scope attached. Hidden assumptions are what a side-by-side review is meant to remove: which element changed, which state was in view, and what was explicitly rejected. Implementation, tests, and any repository write remain your work. A keep in the review is not a commit.
When a screenshot thread is the wrong tool
A markup on one frame is fine for a narrow visual note. It gets weak when the group has to agree on behavior or on what will be built. The screenshot comparison is the longer version of that distinction. If the proposal came from a generator, start from the AI-generated UI review instead, and still hold it against the current product.
Questions
Do product, design, and engineering need separate artifacts?
No. They look at the same Current and Proposed comparison and the same decision record. The record carries the scope, the preview, and the next step. It does not replace the implementation work.
What if I am reviewing alone?
Start alone. Try That UI is for individual developers, designers, product managers, and founders as well as teams. Share the same comparison when another person needs to judge the change.
Does the review ship the change?
No. Keeping a direction records the decision. Code, tests, and any repository write stay outside the review unless a separate approval is given for that operation.
Create a free account