Design Process

Preface:

We at Grey Chapel have 10+ years of graphic design and digital image editing skills that we bring with us into our designs and our process. We've done all sorts of things before t-shirts under different names, such as custom packaged trading cards, all manner of signage and banners, even stuff as simple as business cards and invitations. We are also programmers with just as many (if not more) years of coding and other IT-related projects under our belts. That experience has culminated and aided in the creation of Grey Chapel as you see it today.

Of the many tools we use in our workflow, generative AI is the one that we get asked about the most. We use it very early in the process to rapidly test ideas and produce a starting base image, much like one might sketch drafts or pull references. Below is our full design process, including exactly where AI does and doesn't show up.

The Design Process

We start with an idea. Design ideas typically come from various places of inspiration and sometimes from our market data. We do not use AI to come up with the ideas for the designs, and we didn't use AI to come up with the Grey Chapel style for our designs. 

We use AI to quickly iterate through our design ideas, seeing if they'll work out or if we need to go back to the drawing board. We perceive the use of AI to be another tool in the digital process, much like Photoshop is. If we get a result that we think looks nice (or can with some fixes), then we will typically use that image as a starting point for a design.

Once we have a base image, we get to work. Everything from this point forward is handled by us, using tools and techniques that have been common in the industry for as long as it's been around.

With just about every generated image that makes it through our approval process, there are issues with it that prevent it from being used directly. These will be things such as AI artifacts, inaccurate or missing details, or the whole image being not optimal for conversion into a graphic at all.

We fix these issues by hand using image editing software such as Photoshop or Affinity. We could use AI to fix some of these problems, but more often than not the AI will either misunderstand what we actually want fixed, or it will introduce other issues that are not possible to fix. In this case, fixing them by hand is the faster and usually better choice.

Sometimes the AI will produce an image or design that looks awesome, but it wouldn't be possible to fix the image into a usable design without extensive work. More often than not, we are forced to abandon these finicky designs and move on for sanity's sake.

Once an image makes it through the fixes and polishing phase, it gets another QA check, and then it gets vectorized. We have an internal tool that automatically traces the design from the image, giving us a layered set of vector graphics from our refined images.

These vector graphics then get their own wave of cleaning up, where we either simplify, move, or outright delete various graphics in order to bring them up to our vision for the finished design.

Then we set it up as a print-ready file, and we're done. We have a finished Grey Chapel t-shirt graphic. This design will be uploaded to our store, typically in batches, for you to purchase on your preferred shirt color and size.

It's also important to note that we don't use AI for every design. Some designs are made using images not generated by AI, some use elements generated by AI (but are otherwise composed by hand), and some are made entirely from scratch digitally. It depends on the design and our final vision for it.

So, to recap, in the creative/design process:

What we use AI for:

  • Creating the starting or reference image for most designs, or specific elements for others

What we do not use AI for:

  • Creating design ideas
  • Creating design styles
  • Editing or fixing issues with the starting image
  • Vectorizing the design
  • Cleaning up the vector graphics
  • Producing the final print-ready file

The Mockup Process

The same sort of process applies to our mockups, but it's a little bit simpler. The process for mockups looks like this:

  • First, we create a base image of a blank t-shirt. This includes the shirt, the background, and the set dressing, using generative AI.
  • We will then check the image to ensure that it's not misleading, incorrect, or misrepresentative of the actual t-shirt we are selling.
  • At this point, we may use AI to generate different variations of the shirt, like different color variations, test different backdrops, etc... It's far simpler and faster to edit the images in this way using AI than with traditional methods.
  • Once we're happy with it, we clean up the image by fixing any AI artifacts, visual glitches, color mismatches, etc... all by hand, by us.
  • Last, we'll detail the image. This step is adding details back into the image using traditional photo editing techniques to make the image better represent our intentional vision, and better represent our products. Again, all by hand, by us.

Now we have a base image that's ready for a design to be placed on it. From here, we have an internal tool which automates bulk creation of the mockups for every color we offer by placing a design on the mockup, making it blend it in with the shirt as it would when printed, and overall make it look like the graphic was on the shirt from the start.

For mockups, we are basically forced to use traditional methods to place the designs onto the shirts, otherwise the image may be misrepresentative of what you receive. AI, while good, tends to make mistakes about design details and such. Meaning that even if you hand AI the exact design and shirt, anything from the colors, the size/placement, and even the design itself can become distorted or otherwise misrepresentative of what you will actually receive. AI not working great for this purpose doesn't bother us though, because we actually like doing it.

 

We hope this breakdown of our design and mockup processes has cleared up any questions you might have. If you have any questions or comments, feel free to use our contact form to send them our way.

Last updated: 09-21-2026.