Adobe Firefly didn't arrive in a vacuum. By the time it showed up inside Photoshop and Illustrator, tools like Midjourney, DALL-E, and Stable Diffusion had already proved one thing: you could turn plain text into visuals that were good enough to influence real client work. Firefly's big difference is where it lives. Instead of being "another app," it's baked into the tools many designers already use every day, which makes experimenting feel less like a separate workflow and more like a natural extension of the creative process.

And that's the real story here: Firefly isn't about replacing designers. It's about speeding up the boring parts, opening up faster experimentation, and giving you more options earlier—while the designer still decides what's right.

What designers actually use day to day

Two designers described Firefly as something that quietly slips into routine work rather than a dramatic "AI revolution moment."

Photoshop: Generative Fill and Generative Expand

One designer leans on Generative Fill almost daily, especially when a photo needs "one specific thing" added, removed, or adjusted. Generative Expand also comes up a lot when an image needs extra breathing room for layout—like stretching backgrounds to fit new aspect ratios without the usual cloning and patching marathon.

Illustrator: Text to Vector Graphic and Generative Recolor

On the vector side, Text to Vector Graphic is becoming a handy way to generate starter assets from prompts, while Generative Recolor helps explore multiple palettes quickly—especially useful when clients want "more options" but can't describe what they mean beyond vibes.

Why this helps: the stock image problem

If you've ever dealt with stock sites under deadline, you already know the pain: clients want something weirdly specific, and stock libraries give you "almost, but not quite."

One designer described the classic scenario: the client wants a particular person, holding particular items, standing in a particular place, wearing a particular style, with specific color accents. The old approach was messy: find multiple stock photos, pay for them, then stitch them together with heavy edits—and still end up with something that looks slightly off.

With Firefly, the workflow changes. You can start with one "close enough" base image and use Generative Fill to push it toward the client's request. The big win isn't perfection—it's time. What used to eat up hours of searching and compositing can become something closer to an hour including the research.

A quiet game-changer: icons and custom assets

Icons are another sneaky time sink. Traditionally, you hunt through libraries, then tweak a symbol until it matches your need, and sometimes deal with licensing depending on the source.

With AI-assisted generation, designers can create icons that are more specific and less "I've seen this exact icon on 40 other websites." It also helps build a more unique visual language, especially for brands that want to feel distinct without paying for a fully custom illustration system upfront.

Is it beginner-friendly? Pretty much, yes

A recurring point was how "obvious" Adobe made these features. The AI options aren't buried in hidden menus. They show up when you're likely to need them, and they don't require you to learn a whole new interface.

For beginners, that's a big deal. If you can select an area and type what you want, you can start. And if you don't know what to type, some tools will still offer suggestions—so even experimenting feels approachable.

Where Firefly still falls short

The designers weren't pretending it's magic. They pointed out a few practical limitations that show up quickly in real work.

The "not quite Midjourney" feeling

For raw creativity and polished concept imagery, some designers still prefer tools like Midjourney (or other generators) to create the initial idea, then bring the result into Photoshop to refine. Firefly becomes the finishing tool rather than the spark.

Editing AI images can be oddly harder

A regular photograph has consistent lighting, texture, and logic. AI-generated photorealistic images can have tiny inconsistencies that make editing frustrating—so instead of "fix it once," you end up regenerating variations until something finally behaves.

Vector generation needs cleanup

On the Illustrator side, AI vectors can come out looking slightly "assembled" rather than designed—like pieces joined together instead of clean, intentional shapes. Prompts help, but they don't eliminate the need for human polishing. In practice, it's a starting point, not a final logo-ready result.

What UX designers wish Adobe would automate

One of the most interesting takes came from the UX perspective: UI work has a lot of repetitive setup that isn't truly "creative," just necessary.

Think of message lists, avatar grids, galleries, placeholder names, sample content, and repeated components. The wish is simple: let AI recognize the pattern and fill it in intelligently—faces in avatar circles, realistic names, believable snippet text—so designers can focus on structure and experience instead of busy work.

Exploring ideas faster without losing control

Firefly's "multiple options at once" approach is a big part of why it helps ideation. Instead of generating one result and hoping it's right, you get variations. That makes it easier to compare directions quickly and choose what fits before you invest time in refining.

One designer even described using the client's own wording as the prompt for recoloring a brand palette—and getting results closer to what the client actually meant. That's a pretty realistic use case: clients often know what they want emotionally, but struggle to translate it into design language. AI can bridge that gap by generating options you can react to.

Keeping the human touch (so it doesn't all look "AI-ish")

Both designers landed on the same mindset: start with your own idea, then use AI as a tool to execute faster—not as the thing deciding the direction.

A practical way to keep control is to "judge AI output like you'd judge a human designer's draft." When you do that, you'll notice what still needs improvement—especially typography, composition nuance, and expressive detail. The designer's taste is still the filter that matters.

Practical advice if you want to use Firefly well

Final thoughts 

Adobe Firefly is most powerful when you stop thinking of it as "AI art" and start treating it like a productivity layer inside tools you already know. It speeds up searching, expands options, helps you iterate, and reduces the grind of repetitive asset work—but it still relies on your taste, your judgement, and your understanding of the audience. Used that way, Firefly doesn't shrink creativity. It gives you more room to use it where it matters.