A year ago, designers were still arguing about whether AI was a threat or a tool. That conversation is over. AI is here, it’s in your workflow already, and the question now is which tools are worth your time.

Image generation is getting better

The image generation space has settled into a few clear leaders. Midjourney still produces the most striking visuals, but you need to know how to talk to it. Writing prompts that get you what you want takes practice. DALL-E 3 through ChatGPT Plus is more forgiving, it understands messier prompts and interprets what you probably meant. For concept work and mood boards, both save hours compared to stock photo hunting. Google’s Imagen 2 caught up recently with better photorealism and fewer of those telltale AI quirks. It handles text in images better than most, which matters if you’re mocking up posters or packaging. The catch is you need it through Google’s platforms, so it’s less convenient if you’re working in Adobe all day.

Adobe built it right in

Adobe built AI directly into its tools, and that integration is hard to beat for working designers. Photoshop’s Generative Fill does exactly what it sounds like, you select an area, describe what should be there, and it fills it. The results are good enough for client work about 60% of the time. The other 40% needs tweaking, but you’re still faster than doing it manually. Illustrator added generative AI for vector graphics. You can generate icons, patterns, and illustrations that are immediately editable as vectors. This is huge for logo design and branding work where everything needs to scale. The style control isn’t perfect yet, but it’s getting better with each update.

Tools for UI and UX work

For designers who don’t live in Adobe Creative Suite, Figma added AI features that make sense for UI/UX work. The AI can suggest layouts, generate placeholder content that looks real, and auto-organize layers. These aren’t flashy features, but they cut down on the tedious parts of interface design.

What’s happening with logos

Logo design got interesting with tools like Looka and LogoAI. They’re not replacing designers, but they’re good enough that small businesses use them instead of hiring someone for basic work. If you’re a designer, this means your logo projects need to offer more than just a mark, strategy, applications, brand guidelines, the full system.

Photo editing on autopilot

Photo editing has AI baked in everywhere now. Luminar Neo and Topaz Photo AI handle batch editing, sky replacements, and portrait retouching faster than manual work. For product photography and ecommerce brands, these tools mean one designer can handle volumes that used to need a whole team.

3D without the learning curve

The 3D design space is where things get weird. Spline added AI generation for 3D objects, and it’s surprisingly usable for web graphics and simple product mockups. You describe an object, it generates it in 3D, and you can rotate and customize it. The quality isn’t there for complex work, but for web design and quick concepts, it’s faster than modeling from scratch.

Changing motion design

Text-to-video tools like Runway and Pika are changing motion design. You can generate short video clips, extend footage, or change elements in existing videos. For social media content and quick animations, these tools mean designers don’t need to be video editors anymore. The transitions still look off sometimes, but clients often don’t notice.

An all-in-one approach

Canva pushed hard into AI this year with their Magic Studio features. Background removal, image expansion, text-to-image, even a presentation generator. For freelancers juggling multiple clients, having all this in one tool means less app-switching. The quality doesn’t match specialist tools, but the convenience factor is high.

The small tools

Color palette generation got smarter. Tools like Khroma and Huemint use AI to suggest color schemes based on your preferences. They learn what you like and generate palettes you’ll probably use. This sounds minor, but color selection eats up more time than it should, and these tools speed it up. Font pairing is another spot where AI helps. Fontjoy generates combinations that work together, which is helpful when you’re stuck or working fast. Monotype added AI to their font platform that suggests typefaces based on your project description. Not groundbreaking, but useful when you’re browsing thousands of options.

Single-purpose tools

The smart object and plugin space exploded. Remove.bg still does background removal better than most built-in tools. Cleanup.pictures removes unwanted objects from photos cleanly. These single-purpose AI tools do one thing well, and you can string them together into a workflow that’s faster than doing everything in one app.

Scaling up for agencies

For agencies, the real value is in the AI tools that handle repetitive work. Bulk image resizing, format conversion, generating variations of a design, this is where AI saves the most time. One designer with the right AI tools can output what used to take three people. The personalization tools matter for marketing work. Recraft and Leonardo AI let you generate brand-consistent visuals at scale. You train them on your brand’s style, then generate hundreds of variations for different campaigns. For brands running lots of A/B tests or personalized content, this changes the economics of design.

Where AI still struggles

What doesn’t work yet: AI still struggles with hands, complex text layouts, and maintaining perfect consistency across a series. If your project needs those things, you’re still doing it yourself. And AI-generated work has a sameness to it that sharp eyes can spot. Original concepts still come from human designers.

The workflow

The workflow that makes sense for most designers in 2025: use AI for the grunt work, ideation, and variations. Use your skills for the creative direction, refinement, and the touches that make work distinctive. The designers who figured this out are billing more because they’re faster, not because they’re working harder. Free tools exist for almost everything mentioned here. The paid versions are faster and have better outputs, but you can test the workflow without spending money. Most designers end up with two or three paid AI tools and a handful of free ones, depending on what kind of work they do most.