The Membership Blast — 028

How We Save Hours Using AI Thumbnails (Steal This)

* indicates required

Hey from the Northside!

Spring showed up early in Chicago this year. I’ve already been to a couple of Cubs games, and it finally feels like we’re out of hibernation.
 
Alright, onto something that’s been saving us a ton of time lately…

How we use AI Projects to scale thumbnails (without starting from scratch every time)
 
If you’re creating content regularly, course images, community posts, YouTube thumbnails, etc, you know the grind.
 
  • Open a design tool
  • Try to remember your style
  • Recreate it… again
  • Tweak colors, fonts, layout
It adds up fast.
 

What we started doing instead is using AI Projects inside ChatGPT to standardize everything.

The simple workflow
 
1. Create a new Project in ChatGPT

This becomes your “design hub” for thumbnails.

2. Add a master prompt that defines your style
Here’s a real example you can adapt:

Create a thumbnail image with a clean, cartoonish illustration style.

Style:
– Slightly exaggerated cartoon version of a person
– Bold outlines, simple shading
– High contrast and visually clear at small sizes

Brand:
– Primary color: #4F46E5
– Secondary color: #22C55E
– Background should be simple and uncluttered

Typography:
– Bold sans-serif font
– Large, easy to read
– Minimal words (3–6 max)

Layout:
– Subject on one side
– Text on the other
– Include subtle visual elements related to the topic

Tone:
– Friendly, modern, slightly playful
– Designed for creators and entrepreneurs

Always keep consistency across all images.

3. Upload 1–2 sample images (Optional)
If you already have thumbnails you like, upload them to the project. Now the AI has visual context, not just instructions.

4. Test your prompt
In the same Project chat, ask something like:
“Create a thumbnail for: ‘How to Sell Your First Online Course’”
Refine until it feels right.

5. Reuse it forever
Now whenever you need a new thumbnail, you just say:
“Create a thumbnail for: [your topic]”
That’s it.

No re-explaining your style. No starting over.

Why this works so well

Two big reasons:

1. Consistency without effort
Your brand stays tight across every image automatically.

2. Cartoon styles scale better than realism

IMO, AI still struggles with creating realistic images of headshots, especially consistency across images.

Cartoon or illustrated styles solve that problem. They’re easier to generate, look clean, and actually stand out more in feeds.

That’s what we did for our newsletter.

What this replaces

Before:
  • 20–40 minutes per thumbnail
  • Constant back-and-forth tweaking
  • Inconsistent style
After:
  • 1–3 minutes
  • One prompt
  • Consistent output every time
If you’re publishing even a few pieces of content per week, this can save hours.

If you want to use real people

You still can, just be more specific:

Use a realistic photo style with studio lighting.
Keep the same person across all images.
Neutral background, soft shadows.
Consistent framing (chest up).

Just know it may take more iterations.

What we do at MemberSpace

We use a Project like this for our newsletter and content.

When we need a new image, we just ask for it. The system already knows our colors, style, and layout.

See it live →
AI thumbnails example

Our newsletter thumbnails and content live inside MemberSpace. If you’re not a MemberSpace customer, create a free account to view the thumbnails or leave us a comment.

It’s one of those small changes that quietly compounds over time.

If you’re creating any kind of content regularly, this is worth setting up once.

You’ll feel the difference immediately.
 
Talk soon,
 

Marvin
Chief Growth Officer
MemberSpace