A little
context.
The Navy needed a quicker, more efficient way to build and evaluate daily aircraft mission plans, known as Air Tasking Orders. An AI tool was already in development to generate and score possible plans. Our role was to make that tool easier to use.
I worked with two other designers, developers, and human factors engineers. We had about three months to design a high-fidelity prototype, with limited access to end users. Throughout the project, we shared our work with the Air Force Research Laboratory for feedback.
There was a lot to learn before designing.
The planning process was complex, and the team needed a shared understanding of how it worked. We met with subject matter experts to learn about the domain, the goals of the planners, and the problems in their current workflow.
Frequent manual updates left planners with limited time to compare alternatives. We needed to make the existing prototype clearer while still supporting the detail involved in evaluating a plan.
Functional decomposition
Understand the high-level goals and sub-goals.
Information architecture
Decide where the features should live.
Concept maps
Show how the parts of the tool connect.
I started with sketches to get feedback quickly.
Once I understood the problem and the goals more clearly, I began exploring ideas through low-fidelity sketches and wireframes. This let me try different approaches without spending too much time polishing them.
I presented the concepts to the team and gathered feedback from subject matter experts. After making adjustments, I moved the strongest directions into high-fidelity mockups.
The experience needed to work across two screens.
Our research showed that planners typically worked with two monitors. That became an important part of the design: they needed to compare plans, then look more closely at how one would play out.
The comparison view showed the top three plans selected by the AI and their metrics. Selecting a plan revealed more detail, with the map and timeline on the second monitor.
Compare the options
Review the top three plans and their metrics.
Explore one plan
Use the map and timeline to understand the selected plan.
From comparing a plan to making adjustments.
The designs also needed to support what happened after a planner chose an option. They could adjust selected metrics, click into elements for more information, and save a revised plan to compare with the others.
I also designed a presentation mode that brought the table, chart, and metrics onto one screen. This gave the team a way to present the final plan in a different context from the two-monitor working setup.
Compare
Review the available options.
Refine
Explore details and adjust metrics.
Save & present
Compare the revision and present the plan.
Keeping the prototype consistent as it evolved.
The NIWIC team had a design system they wanted us to use. I incorporated its colors, typography, components, and icons into the final designs, and updated the existing UI kit as the prototype changed.
We met regularly to critique the work and shared the high-fidelity prototypes with AFRL for feedback. The work shown here focuses on that prototype and the decisions behind the planning experience.