Spring 2023
Interface Design, Artificial Intelligence, User Experience


This project is an exploration of how AI technologies, now and in the future, could be applied to the design of an interface meant to support the translation of foreign intelligence by analysts.

Through our partnership with LAS, we were able to gain valuable knowledge and feedback throughout the process. One of the largest hurdles in this process is that the software currently used by analysts is classified. This meant we only had a rough idea of what their current interfaces look like and how they function.

The goal of our design was to aid the user in creating robust and reliable intelligence that conveys content, intent, and context. The following questions guided our investigation: 

  • How can the user not only translate the intelligence accurately but also learn the context of the situation and the people involved?
  • How can the system support continuous learning of foreign languages through grammar and word choice?
  • How can information gathered by the analyst and system be filtered for faster scanning?
  • How can the system intelligently fix an errorful speech-to-text, as well as make it easy for the user to make those changes manually?

Important Note: For the sake of presenting real data the transcripts, names, and events are adapted from the Nixon tapes. The scenario dictated that these transcripts were coming from a foreign intelligence in another language and being translated into English.

Meet Cameron

Cameron is a Full Performance Language Analyst who has just been assigned to a new team.
This project made use of a human-centered design process to create an interface experience that would support specific personas provided to each team. My team, comprised of myself and Ned Babbott, was given the persona of Cameron, a novice translator.

Cameron’s experience, strengths, and weaknesses laid the groundwork for the kinds of problems we would tackle with our interface. For example, while Cameron has been a Language Analyst for three years, they have almost no experience with the language of the new country they have been assigned. Our system needed to support accurate translation as well as learning opportunities while Cameron was working.
Below you can find our user journey maps. These maps helped us pinpoint areas where Cameron struggled throughout the day. For us, a successful system would eliminate most, if not all, of Cameron’s challenges that stem from the systems they use.

From what we were able to learn from our stakeholders, actual language analysts, there are too many tools they need to use that don’t work well together. We illustrated this in the maps by plotting out what tools Cameron might access at a given time.

As Is User Journey
To Be User Journey

More Research

As mentioned, this project gave us access to real language analysts through our partners at LAS. We met with the team periodically throughout the project to ask them questions and get feedback on our work. 

One of the first interactions was a formal interview with individual members of the team so we could gain a better idea of what it was like to be an analyst. For many of us, this was our first time even considering the fact that language analysis was a career. While we couldn’t become experts in such a short time period, we needed to get close in order to create accurate design solutions.

The interactions that followed were formal critiques of the concepts we were developing. While having access to experts was extremely helpful, we also had to remain focused and not let an influx of ideas change our direction too much. We had to communicate clearly and effectively with the analysts to make sure we made the most of our time with them.

Now, Design

The design of the interface began with three concepts based on resolving different pain points. The initial concepts were related to correcting an errorful speech-to-text, visualizing contextual searches, and mapping storylines.

Feedback from analysts allowed us to focus on which issues we wanted to tackle and how the different tools might work together within a unified workspace.

The visual language was developed to enhance the focus of the analyst within the workspace. Dark colors around the edges would draw the eye to the transcription space in the center. Bold colors would alert the analyst of problems or new information. Finally, simple typography would keep the text readable throughout.
Initial ideas as rough wireframes
Narrowing down, sketching, and developing the visual language
Context builder and grammar tool hi-fi sketch

Let’s Walk Through It

The final sketches and user journey were combined to create a walkthrough video.

Credit to Ned Babbott for animation and voiceover work.

A Little More

Finally, here are some of the key features of the system.

This project was larger and more complex than I could have ever expected. I learned a lot not only about design but also about a whole field of career and study I had never been exposed to before. A huge thanks to everyone at LAS for the support throughout and being our biggest hype-people.

This project was later presented to the wider United States intelligence community to showcase how involving designers in unexpected places can yield well-designed applications of future-thinking technology.
Copyright Kevin Ward 2023 ©