AI Conversation Analytics Tool
Background
After my application was accepted for an AI startup, I was tasked with completing a test assignment to demonstrate my skills and vision for their product. Due to by background in human behavior, I was drawn to this company's goal and felt I would be able to enhance their product with an understanding of how to understand to others and how people intake new information.
Research
Discovering Needs
Without the availability of user interviews, I began by diving into what problems had already been discovered in AI Analytic tools. Following this, I scrutinized the current interface to uncover pain points for usability.
How can I make this informative interface more digestible to users?
What changes facilitate a higher level of accessibility?
How will users interact with this information?
Market Research
AI tools experienced a sudden boom from 2022-2024. By diving into what barriers other companies had faced in the creation and usability of these tools, I found a few key areas that I could account for in my design.
Analysis
As this project occurred over a one week time period, I had to find a way to research without conducting user interviews. Focusing in on each screen that the company provided me with, I analyzed what problems a user could have while navigating the information.
Key Takeaways
The analysis provided by the AI needs to include an explanation for the information to be usable.
The UI for this product can be improved to show a heirarchy of information as well as undergo changes to enhance accessibility.
Conversations need to be easily editable to ensure that the data being used is accurate.
The AI system needs to understand the user context for quality analytic output.
Ideating Solutions
My Approach
Given the prompt for this project, I decided to focus on UI design changes that could be made to enhance their current flow.
How can I make this informative interface more digestible to users?
What changes facilitate a higher level of accessibility?
How will users interact with this information?
Sketches
With a tight deadline and many possibilities, I began by sketching and narrating my ideas, testing different layouts before diving into higher fidelity.
Styling for Accessibility
In order to give a basis for designing with accessibility in mind, I created a style guide that kept the company's current blue aesthetic while ensuring that backgrounds and text passed accessibility checks within this theme.
Redesigning UI
Designing the Interface
Approaching the UI, I focused on creating a visual language that could be easily understood and navigated by potential users.
Information architecture is key for informative input.
Accessible colors and spacing allows for all users to access the information.
Interactive tabs allow for a large amount of information to be shown on one page, minimizing user navigation.
Changing the Layout
Focusing on the visual way that users would interact with this information, I redesigned the interactions and layout for each section that the startup prioritized in their analysis. Within this redesign, I made sure to create room for reasoning to be provided by the AI, giving users a deeper understanding of why their relationship to the other person was ranked at a certain level.
Enhancing Accuracy
To give the user control over the accuracy of the information being fed into their analytics, I redesigned the transcript to be an interactive conversation similar to a group chat. This allows the user to begin creating profiles for different speakers and ensure that the conversation data is accurate.
Current Interface
Redesign
Prototyping
Moving Forward
Pausing Design
The company has informed me that they have paused the redesign of this product and are working on developing their current interface. I am still in the pool of applicants should they revisit this project.
Future Ideation
As a part of my design submission, I explored what key metrics would be important to track in the future and what changes in flow could enhance the product. I also debated the legal implications for a product that relies on recording conversations for producing output.