Build Your Business:


Integrating AI Into Your Company’s Custom Apparel Design Process

Or how I learned to stop worrying and love the algorithm when design art for T-shirts, hoodies and other apparel for screen-printing, DTG, DTF and other decorating methods

By Marshall Atkinson, Contributing Writer


An example of the many results you will obtain after entering a series of prompts into a generative AI, program. The program in question in this case is Midjourney. Image courtesy of Marshall Atkinson

May 9, 2024

Can you feel it? There is a revolution underway in the decorated-apparel industry. This isn’t about artificial intelligence, or “AI” taking anyone’s jobs, but about harnessing today’s latest tools to unleash your already existing creativity, streamline design processes and redefine profitability in custom apparel.

Love it or hate it, artificial intelligence is making deep headways into the decorated-apparel industry. AI presents an unprecedented opportunity to level the playing field for small to medium-sized businesses. New tools allow anyone to compete with larger players by offering innovative designs at a fraction of the time and cost.

That said, not everyone is adopting or experimenting with these tools yet. It is a time issue for some: “I don’t have time to play with it.”

For others, their hesitation is based on jealousy or fear. “I’m an artist, there is no way I’m using those tools,” they say.


Still others worry about what they believe to be the ethical challenge of these kinds of technologies. “AI steals artists’ work!” they exclaim.

This article will explore the facets of integrating AI into a shop’s design process. It will also cover these concerns and more, so you can truly understand what’s out there and decide what to do.

Understanding AI in Custom-Apparel Design

At the heart of this revolution are a number of “generative” AI platforms, like DALL-E (openai.com/research/dall-e), Adobe Firefly (adobe.com/products/firefly.html), Ideogram (ideogram.ai) and Midjourney (midjourney.com). Similar to ChatGPT, these platforms all employ something called Large Language Model (LLM) machine learning to analyze gazillions of images and the captions accompanying them. The platforms then “learn” to associate various different words with various different visual concepts. This is how when you type “horse,” “blue” and “cartoon” into one of the programs, you get that tool’s version of a blue cartoon horse.

Generative AI image from Midjourney

A closeup of one of the dozens of Midjourney options shown in the image at the top of this page. Image courtesy of Marshall Atkinson

Contrary to popular belief, these kinds of AI tools do not simply copy and paste a mishmash of related images to come up with a new one. Instead, with each new image request, an entirely new image is created via a multi-step process known as diffusion. Because Adobe Firefly, Ideogram and Midjourney are all trained using different sets of images, you will not only inevitably obtain different results using the exact same prompts.

In addition, because these platforms’ diffusion models all begin with random noise, every image they create will be unique. Results cannot be repeated.

Embrace The Awkwardness of Generative AI

Creative professionals have long mastered various commercial design tools. Photoshop, Illustrator, Corel Draw and other programs have formed the backbone of this industry since their emergence in the early 1990s. Every trained designer knows how to use layers, channels, the pen tool, blur commands and the other graphic toolsets that are part and parcel of these programs.

AI image creation is different. Suddenly, your success has as much to do with your vocabulary as anything else. Professional designers who have mastered these skills are suddenly being out designed by that accountant in the corner. It feels unnatural. No wonder there is some friction within parts of the design community.

The good news is an untrained AI creation tool user doesn’t have the training, knowledge or creative mind to utilize these tools fully. A professional designer using these new tools will invariably create and iterate better images than someone without that kind of creative eye—it’s only natural. They simply need to get used to how these platforms operate.

Which is the weirdest thing you can imagine. Words.

As they say, words matter, and with AI design tools even more so. The fun part is the results stemming from different words that mean the same thing can produce different results. For example, the word “minimalistic” can be interchanged with “spartan,” “simple,” “sparse,” “austere,” “basic” or “unique. “Each word will affect the results differently when inserted into the same prompt sequence.

Words and word sequences are the alchemy employed to create images as opposed to the tools the profession is used to. New images can be rendered in under a minute. On Midjourney, you can simultaneously create several dozen images with the correct prompt commands in little more than the blink of an eye.

Choosing the Right Generative AI Platform

Everyone has different tastes, skills, work habits and budgets, and each of the aforementioned AI tools works a little differently. Here’s a breakdown of what you need to know:

DALL-E 3

This platform is owned by OpenAI, the same outfit that brought you ChatGPT. Since you will be using images created for print, an HD image measuring 1024-by-1024 pixels will cost you $0.08, while a 1024-by-1792 or 1792-by-1024-pixel image will run you $0.12. DALL-E 3 is trained on 400 million image pairs. You can use DALL-E 3 for free on the Bing Image Creator.

Adobe Firefly

This is baked into the Adobe suite, which many graphic professionals already use. If that’s you, you’re already paying for it. There is a free plan, but users who upgrade to the premium version ($4.99 per month or $49.99 per year) can remove the Adobe Firefly watermark. Adobe Firefly is trained on the images in the Adobe stock image universe, which occasionally makes the results look somewhat limited. Adobe Firefly is the backbone of the generative fill engine for the Photoshop upgrade.

Ideogram

There are multiple programs/price points on offer from Ideogram. There is a free plan that gives you 100 standard images a month; a basic plan for $8 a month ($84 annually), giving you 1,600 priority images monthly; and a Plus Plan for $20 a month ($192 annually) that gives you 4,000 priority images monthly. The data set information for Ideogram is not available.

Midjourney

This platform has four price points. The Basic Plan is $10 per month ($96 annually) for unlimited images, but 3.3 hours per month of fast GPU time. The Standard Plan is $30 per month ($299 annually) with unlimited images and 15 monthly hours of fast GPU time. The Pro Plan is $60 per month ($576 annually) with unlimited images and 30 hours of fast GPU time per month. Finally, the Mega Plan is $120 per month ($1,152 annually) for unlimited images and 60 hours of fast GPU time monthly. Midjourney has yet to disclose how many images have been used to train its model, but it has been reported to be in the hundreds of millions if not over a billion images.

Like a comfortable pair of blue jeans, you should try on all four of these platforms to see which is the best fit. Personally, I’ve found that DALL-E 3 and Midjourney provide the best images, while Ideogram is the first platform to master text embedded in the design consistently. There is no wrong answer.

Full disclosure, it’s no secret that I have a serious bromance with Midjourney. I find it’s faster than the others and that the images created on the platform are better. As a professional designer, I’ll add text on a different platform anyway, so most of the time that isn’t an issue. That said, I play with all four platforms constantly, as I want to learn how to get better results.

Overcoming the Design Learning Curve

These tools can be intimidating. You can be gloriously happy with the results one minute and deeply frustrated the next. How to get better results is not always obvious. What follows is a list of some tips you can deploy when playing with these tools:

  • Create a spreadsheet of the prompts you’ve tried and what they do. Art, photography and printing vocabulary words go a long way with these tools.
  • Don’t be afraid to mix and match words in prompt sequences. Even non-logical ideas can produce a result, sometimes an amazingly good one.
  • Play like a child with a new box of crayons. Just create.
  • Google and a good thesaurus are your new best friends. Look up words and try them.
  • The best images usually come from remixing/editing the results and pushing the images. You will often get good results with the first attempt. However, four or five remixes or some additional prompt edits can deliver results that are nothing less than amazing.
  • Set aside five or 10 minutes every day to try at least one new thing. Consistent practice will deliver an understanding of the tools. Try to create images you will use in your work.

There are plenty of resources out there to help you. Watch videos, read newsletters and articles or share ideas with colleagues or friends.

Cost-Benefit Analysis of Generative AI for Apparel Decoration

In addition to helping spur greater creativity, AI tools can help you become more profitable. These platforms significantly reduce the time required to generate artwork, allowing your creative team to experiment and innovate. This efficiency can translate into cost savings, higher throughput and the ability to produce that much more work in a given day.

On the sales and marketing side, AI can allow your creative team to create better images as well.

Without AI, a single elaborate design can take hours to create, whereas that same job might be whittled down to 15 or 20 minutes with AI. Next thing you know, those same projects and ideas your shop never seems to be able to get to in a timely manner are already slotted into your design schedule.

What’s Possible in Terms of Design

Although drawings, graphics, designs and logos represent the most obvious kinds of designs that can be achieved, these kinds of AI tools can be used for virtually any other image or collection of images you can dream up as well. Think background textures, like wood, stone, glass or concrete; useful creative nuggets like camouflage or embroidered flowers; or even simple bandana patterns.

Do you need a doodad thingamabob to fill in a corner or add a bit of funky weirdness to a design? You’re now only a few words away from something that might fit into whatever is floating around in your head or doodled on a post-it note. A set of alien icons? A plastic bobblehead lion in a football uniform? A 1950s beach poster? Pretty much anything and everything is possible with the right prompt words.

Beware of Using of Companies’ Trademarked Material

Uncle Ben from Spiderman was right. Just because you can do something doesn’t mean you should. Yes, you can use any artist, corporation, public figure, logo, brandmark or idea as the basis for your prompt sequence. The question, though, becomes, “Do you own that or have you licensed the rights to use that piece of intellectual property?”

Finish t-shirt design generative AI image Midjourney

A finished logo design combining text with the generative AI image shown above. Image courtesy of Marshall Atkinson

No? In that case, slow down, partner! You’re going to want to curb your impulse to create images that could get you into legal trouble. Granted, these tools will imagine and deliver anything matching your prompt words. Micky Mouse? You bet. Mickey Mouse eating a taco with a side of guacamole? Coming right up, all in under a minute.

The problem is this: using these kinds of images without permission is a great way to you get yourself into a world of hurt with an army of flesh-eating lawyers.

Design Workflow Process With Generative AI

Let’s face it: we all use different tools and methods as part of own design processs. Can I write a perfect workflow that matches what you do in your business? Probably not. What follows, though, is a generalized one that may spark an idea for how you can start integrating AI into your work.

Project: Rockdale Blue Knights

Scope: Create a T-shirt design on a black shirt with the Rockdale Blue Knights logo as a left chest and a knight image on the back in an action pose. Add the Rockdale Blue Knights logo in a matching color.

Shirt Color: Black

Screens: (Underbase White, PMS 7702 left chest) + (Underbase White, Cool Gray 4, PMS 7702, White full back)

Workflow

  • Create a thumbnail idea of the Rockdale Blue Knights back print. This will feature a standing knight with a battle axe and a shield in the background, and the Blue Knights logo centered in the foreground.
  • Go to an AI platform, like Midjourney.
  • Enter the prompts: Menacing knight, full plate armor, wicked great axe, giant shield, dynamic action pose, ready for battle, cinematic lighting, fantasy art style, detailed, textured, high contrast, light blue reflections, black and white, isolated, flat black background, —ar 3:4 —stylize 100 —v 6
  • Midjourney produces a series of images. Remix and work the image until something is acceptable. Total time: 2 or 3 minutes.
  • Upscale and download.
  • Open in design software, such as Photoshop, Affinity Photo or Corel Draw.
  • Edit the image to prep for separation. Total time: 10 minutes.
  • Add logo. Flatten and create a demo mockup. Total time: 1 minute.
  • Send to the client for proof.

Total time invested: 14 minutes.

After the client approves, separate the file. Total time: 5 minutes per color. Total time for creation and art separation: 29 minutes.

Bottom line: with the help of an AI image creator, you’ll be able to knock out best-selling art that will make your customers extremely happy. What are you waiting for?

Marshall Atkinson is a veteran designer, custom apparel decorator, business coach and principal of Atkinson Consulting, (atkinsontshirt.com). This past year he launched the online “Midjourney: Elevating Print Creativity Newsletter.