Gabriel Fuller- Portfolio

Description:

LyricalAI

LyricalAI is a prototype interface for a Large Language Model centered around using AI to analyze song lyrics.

Type:

LLM | AI

Interaction Design Studio Group 4

Aug – Dec 2023

Personas

Personas are a helpful tool for understanding how a generalized person may interact with an interface. Two of our personas are Gerald Smith and Alexander Major. These personas will shape how the interface is designed, allowing easier interaction for their needs.

Alexander is a 39 year old Singer/Songwriter who wants to connect with his audience and grow as a musician. He is less technologically savvy, so we must design accordingly.
Gerald is a 22 year old Marketing Manager who uses technology often. He listens to music in his spare time from work, yearning for a greater connection with the songs to provide stress relief.

Storyboards

Storyboards are a way to tell a story for a given persona. Our personas, Gerald and Alexander, each have a reason to use LyricalAI.

Alexander is performing his newest song, while a fan storms off. He uses the interface to understand why the lyrics he was singing were offensive, and change them accordingly. In this interaction, he is able to refine the lyrics and increase his popularity.

Gerald discovers a new song and wants to know what it means. He uses the interface to generate what the “general audience” thinks the song means. By Gerald using the interface in this way, we know that he is searching for a standard understanding of the song and what’s to know it quickly.

Prototype V1

Prototyping was created using Figma. After analyzing our first round of user studies, we had a basic understanding of what was needed of the design.

The interface is split between three panels. The left edge is reserved for clicking buttons to chose which persona you would like to analyze the song. The middle section is where you can paste/type the lyrics you wish to analyze. Finally, the right edge will use Bing for real-time interaction with a LLM.