AI & Automation · Guide
How to use AI to speed up your design work
The useful uses are unglamorous: variations, cutouts, extending images, and the boring production work. Here is where AI earns its place in a design workflow, and where it does not.
The Nextversity teamAI & Automation schoolUpdated August 10, 20265 min read

On this page
The short answer
The genuinely useful design uses of AI are the boring ones. Not "generate a campaign". More like: cut this out, extend that background, give me eight color variations, draft the placeholder copy, clean up this photo.
The pattern worth internalizing: hand over production, keep the decisions. That is where the time savings are real and the quality risk is low.
The tasks worth handing over
Cutouts and masking. Removing a background used to be a fiddly job and is now a click in most tools. Genuine minutes saved, every day.
Extending images. The photo is landscape, the layout needs a square. Generative expand fills the difference convincingly. This alone justifies the feature for a lot of working designers.
Object removal. The bin in the corner, the stray cable, the person in the background.
Variations. Eight palettes, four crops, three lockups. Fast exploration you would not otherwise have time for, then a human choice between them.
Placeholder copy that is not lorem ipsum. Realistic-length text in the right tone makes a layout easier to judge.
Alt text drafts and file naming. Tedious, valuable, and easy to review.
The decisions to keep
What to make. The brief, the concept, the reason this design exists. Hand this over and your work starts looking like everyone else's, because it is drawn from the same average.
Whether it works. Hierarchy, contrast, legibility, whether the eye goes where it should. AI does not evaluate; it generates.
Brand fit. A model does not know that this client's competitor uses that exact blue.
The final craft pass. Kerning, alignment, spacing. The things that separate professional from nearly.
Use AI to speed up the parts you understand. Using it to skip the understanding is how you end up unable to say why the layout is wrong.
Licensing, before client work
Two rules that save awkward conversations:
- Check the tool's terms. Adobe Firefly is positioned around commercial use, with terms published on Adobe's Firefly page. Other tools vary, and the copyright position for AI-generated images is unsettled in many countries. Midjourney's documentation sets out its own position, and the two are not the same.
- Tell the client. Some have policies. Almost none object to AI used for production tasks like masking. Some feel differently about generated imagery in a finished campaign, and it is much better to know that at the briefing stage.
A realistic workflow
- Brief and concept: human. Sketch, reference, decide.
- Exploration: AI-assisted. Generate directions quickly, discard most.
- Production: AI-assisted. Cutouts, extends, cleanup, variations.
- Composition and type: human, in Photoshop, Illustrator or Figma.
- Review: human, and slower than you want to be. Check hands, text, artifacts and anything that looks slightly melted.
That workflow is faster than the old one by a meaningful margin, and nothing in it depends on the AI being right.
Where AI still disappoints
- Text in images. Set type properly. This has not been solved.
- Consistency across a series. Same character, same style, twenty images. Still hard.
- Precise composition. Sometimes drawing it is quicker than describing it four times.
- Vector output. Improving, and rarely production-ready without cleanup.
What this means for your career
The uncomfortable version: production tasks that took hours now take minutes, and some work priced on those hours is worth less.
The useful version: the parts that were always the actual job (deciding what to make, judging whether it works, understanding the client's business) matter more than before. Designers who use these tools are outpacing those who refuse, and the ones doing best are strong at the fundamentals first.
Learn to see, then learn the shortcuts.
Where to learn both halves
The AI & Automation school covers Firefly, Midjourney, Leonardo and Runway. The Graphic Design & Illustration school covers the craft underneath.
One subscription opens both, which is the right way round for this particular pairing.
Hand over the production. Keep the judgment. That is the whole strategy.
Questions people ask
How do designers use AI day to day?
Mostly for production tasks: removing backgrounds, extending images to fit a layout, generating variations, cleaning up photos, and first-draft copy. Less often for finished artwork.
Will AI replace graphic designers?
It is replacing specific tasks rather than the role. Clients still need someone who decides what to make, whether it works, and how it fits a brand. Designers who use the tools are outpacing those who do not.
Can I use AI images in client work?
Check the tool's terms and tell your client. Some tools are built for commercial use and some are not, and some clients have policies about AI content. Getting agreement in writing avoids an awkward conversation later.
Which AI tools should a designer learn first?
Whatever is inside the software you already use, usually Adobe Firefly features in Photoshop and Illustrator. Then a standalone image generator if concept work is part of your job.
Does using AI make my work less original?
It depends what you hand over. Using it for production frees time for the thinking. Using it to decide what to make is where the work starts to look like everyone else's.