In this article, I’ll walk through how I used prompt engineering as part of my design workflow to develop logo concepts for the University of Florida Healthy Gator Coalition. I tried different prompts to provide different directions, generate ideas, and using my design judgment to evaluate the results.

Problem

The Healthy Gator Coalition’s logo was outdated and needed a refresh. The new identity needed to better align with University of Florida’s branding while reflecting the organization’s work supporting the health and well-being of UF staff and faculty.

Prompts

I started by giving ChatGPT a role: a logo designer. I then provided the project brief, target audience, visual direction, and overall tone I wanted to explore.

From there, ChatGPT generated six initial logo concepts in PNG format.

the first set of logo was not exactly what I'm looking for. I then shared a couple visual reference with ChatGPT to have a clearer direction on the concepts because it wasn't aligning with what I had envisioned. Then, I prompted to generate even more concepts with specific directions such as "simple, engaging, heavy on illustration."

Full conversation:

Narrowing down the selection

ChatGPT generated a total of roughly 30 concepts. I narrowed them down to four concepts that best aligned with the project brief. From there, I made my own adjustments to better fit the desired tone and University of Florida’s brand guidelines.

My final logo concepts:

What I've learned:

Using AI as a partner in my design workflow sped up my workflow. The whole conversation took only two hours whereas traditionally, it would take about three to four days.

I'm still learning how to get better at prompting. The more I experiment, the more I learned about how to provide clearer instructions and the best way to direct/partner with the AI tools to solve problems.