AUM® — DESIGNING ONE PRODUCT FOR TEN AI TOOLS

The brief identified what AUM should offer, but left the interaction model open. I needed to determine how people would move from asking for something to working with the result, and which parts of that process could be shared across tools.

I approached the tools through their tasks: entering a request, reviewing an answer, choosing a result, and continuing to edit it. That revealed where one layout could serve several tools and where a dedicated working area was necessary. A document and an image need different controls, but both may need to remain visible while someone gives further instructions.

I kept navigation, account controls, and the position of the model selector consistent across the product. These formed the stable structure around each tool. Inside it, I allowed the layout, input options, and actions to change with the task.

This set a boundary for reuse. Shared elements could be designed together, while a spreadsheet could still behave like a spreadsheet and an image tool could provide visual editing controls. Consistency came from the surrounding structure and placement of common actions.

I proposed three layouts based on how much space a task needed and whether someone was working on one output or two separate tasks.

I did not extend split view to code, documents, presentations, or spreadsheets. Their conversation and working area already occupied both sides of the screen. Supporting another task would introduce more panes and reduce the space available for editing. I chose to preserve that space and limit the available combinations.

Some tools gained additional requirements during the project. As they expanded, I considered whether the new action could stay within the existing view or needed a working area of its own. The workspace layout provided a way to accommodate more involved tasks while retaining the surrounding navigation and controls.

The detailed actions remained specific to each output. A document supports edits to selected passages; an image offers variations and background changes; a video offers actions such as extracting a frame or looping a section. I placed those controls near the result they affect, using the shared layout to organize them.

The scope also included sign-up, login, verification, password reset, light and dark themes, and key mobile screens. These extended the same visual language into entering the product, managing access, and using selected parts of it on a smaller screen.

I delivered the designs for all ten tools and the shared layouts and components supporting them. My scope ended at design handoff; I do not have evidence of how the product performed after delivery.

The central decision was where to draw the boundary between shared structure and tool-specific behavior. Three layouts made that boundary explicit, but their usability still needed validation. The next step would be to test transitions between tools and modes, particularly whether split view provides enough working space and whether people understand when to open a dedicated workspace.