SkyAI.
Wireframes, user flows, visuals, and prototype.
Sky Solutions.
summer 2024.
Lead UX Designer.
Figma, MaterialUI.
Sky Solutions was interested in developing an AI assistant to start marketing to their clients. They wanted something modular, that could be integrated into existing products and client projects. They had no existing AI application, designs, or research so this had to be created from scratch.
Support features requested by stakeholders Create AI interactions to support product Create userflows on how the user would interact with the product Develop basic UI and prototypes
Conducting interviews with several stakeholders including the CIO, I determined that we would need some basic chatbot style functionality. We would also need a way for users to organize and revisit prior queries that were made. This was to be marketed to Sky Solution's federal clients, so having records of what was asked would be necessary. We would also need to allow users to export the generated answers.
"We want an AI assistant to market into our projects"
"Federal users process large amounts of data so they need to be able to pick up conversations where they left off"
Working with stakeholders and developers at Sky Solutions by interviewing them and finding out integral pain points, I started creating a basic user flow for the AI application. I leveraged Material Design and Figma so that development and prototyping times would be faster and we would be creating a consistent style across the entire application. One key takeaway was that federal employees would be working on several items such as cases at once. They would need to inter in queries / have conversations with the AI and then shift to another one. So it was necessary for them to be able to go back and access old conversations and pick up where they left off, adding to them or remixing them as needed.
One of the biggest issues we found is that federal users would need to "go back". Most AI applications on the market do not have this capability right now. Using "cards" to show the previous queries with a short decription helped immensely, prompting users to continue work from where they left off and to ask further probing questions.