CEDIA 2024 Booth
Lead Booth Designer
Snap One
Teaching Reps to Brief AI
Summary
I trained 30 account reps to see why quick AI images look fake, and how our designers write detailed prompts that keep our work looking human.
- Client
- BlueCross BlueShield of South Carolina
- My role
- Art Director
The problem
Some of the internal teams we design for were making their own images with AI, and the results looked obviously AI-made. The cause was simple prompts: a one-line request leaves almost everything up to the AI, while our designers follow a whole process to write detailed ones. Our work also has to feel human, with real, relatable people. That’s why designers stay involved, and why we wanted to stop AI-looking images from reaching the public.
My role
As a member of our AI committee, I built and led a design-led training session for 30 account reps.
The approach
The session showed the reps why those images looked like AI, and how our designers get a better result. It covered the difference between AI image generators and general AI assistants, which tools employees can use at work, and how a detailed prompt changes the result. It showed how repeated edits to an AI image pile up errors, and asked reps to tell the designers which prompt made an image. It ended with a simple guide to the files that work best and the ones to avoid.
The result
Reps loved the training, and the survey afterward was positive. The design team now gets better files, and reps understand why we show the company through real, relatable people rather than an obvious AI look.