Welcome back.

Some tea from Design Twitter:

Henrik from Lovable.dev declared "Webflow is officially dead" after launching their AI tool's visual editor.

Brett from DesignJoy (known for popularizing subscription agencies) wasn't having it.

He challenged Henrik to a live duel:

• Henrik using Lovable.dev​
• Brett using Webflow
• Surprise design prompt
• 45 minutes to compete

Henrik accepted, and I'm hosting the Build Wars this Friday at 11am PT:

• ​Register to watch it for free
• ​Cast your vote here​
• Sponsored by next-gen design tool Paper, and Contra, the marketplace for creative freelancers.

These AI vs traditional tool debates are everywhere in 2025.

Look at what happened with Pentagram getting roasted for using MidJourney while Smith & Diction's AI-enhanced work was being praised the very same day.

Today, let's talk about new tools that are actually shipping design work at a high level.

—Tommy (@designertom)

the wireframe

• The image generation ecosystem in 2025
• Most common tool stacks for prod
• Real-world examples worth studying
• How to use AI image tools responsibly

the reality of AI image tools in design

The debate around AI image tools has been stuck on ethics for too long.

Yes, there are valid concerns about training data, artist compensation, and proper attribution - these aren't going away (and it's worth noting that Adobe - for all their faults - are leading the ethics here).

But while that conversation circles endlessly, forward-thinking designers are showing something more interesting: these tools can be one small piece of a greater creative process, with results no less impressive.​

I've been talking with designers about how they're actually using these tools in production (Smith & Diction shared on Dive Club recently).

The pattern is clear: the best results come from using AI as one component in a thoughtful workflow, not as a one-shot replacement (big reveal, I know).

most common tool stacks for production

After talking to dozens of designers, here's what I'm seeing:

Daily Drivers

​Visual Electric (Free + Paid tier)

Go-to for collaborative brand design:

• Understands design terms
• Brand kit feature eliminates prompt engineering
• Style matching that works
• Infinite canvas for rapid iteration

​Adobe Firefly (Creative Cloud)

Essential for legally safe commercial work:

• Trained only on licensed content
• Generative Fill is a workflow game-changer
• Creative Cloud integration
• Indemnification for commercial projects

​MidJourney ($10/mo)

Unrivaled for artistic concept development:

• Highest aesthetic quality ceiling
• /Describe reverse-engineers art styles
• Strong style consistency with good prompts
• Useless for typography or product specs

Specialized Tools Worth Trying

​RunwayML ($15/mo)

Powerhouse for animated content creation:

• Gen-3 video from images/text
• Frame-perfect motion tracking
• Multi-shot video style transfer
• Green screen replacement via AI matting

​Leonardo.ai ($12/mo)

Champion of narrative visual storytelling:

• Consistent characters across scenes
• Fine detail preservation in illustrations
• Custom model training <1hr
• Best for comic/book artwork

​EverArt.ai ($15/mo)

Corporate brand consistency enforcer:

• Trains AI on product photo libraries
• Batch-generates 1,000+ on-brand visuals
• GDPR-compliant data handling
• Outputs ready for Shopify/Amazon

​Lummi (Free + Pro tier)

Creative visual toolkit:

• Curated AI stock library of 13,000+ vetted images
• Brand color lock for batch edits
• Combines AI generation + human curation
• Style trend presets (photorealistic, doodle, 3D)

​Ideogram ($7/mo)

Typography-first design specialist:

• Embeds flawless text/logos in images
• Handles complex font pairings
• 1920s poster/retro label expertise
• Useless for non-text compositions

TOGETHER WITH BREYTA.AI​

Turn user research into answers, not more work

User research is dope - when you can actually use it. But if you’re spending hours analyzing qualitative data, you’re burning time that could be spent building.

​Breyta’s AI research assistant changes the game by:

• Analyzing multiple user interviews or tests at once (not just single documents)
• Giving you instant themes backed by actual user quotes (no black box BS)
• Creating a searchable repository that surfaces insights when you need them

I dig that it handles the grunt work but keeps you in control. Unlike generic AI that spits out hallucinations, every insight links directly to your data with timestamps.

And they’re serious about privacy: SOC 2 Type II certified, GDPR compliant, European data storage, and most importantly - they never train their models on your research data.

Try it now with a 7-day trial.

​Get instant research answers you can actually trust →​

real-world examples worth studying

Here are designers and teams using these tools effectively:

Brand Imagery (Perplexity)

What works:

• AI for initial concept exploration
• Human refinement of all outputs
• Systematic prompt structure for consistency
• Focus on unique “brand moves” beyond the tool

Content Production (Corona)

The approach:

• One professionally shot hero product
• AI-generated backgrounds for variety
• Human compositing for quality
• 4-day turnaround vs. weeks of location shoots

Brand Photography (DTC Collective)

Their approach:

• Product photography shot conventionally
• AI-generated environmental contexts
• Human-directed compositing
• Enables scale that was previously unaffordable

framework for responsible AI image use

A suggestion for how to generate images:

The Right Way

1. Start with real design work

• Define strategy and core assets manually
• Establish style guidelines first
• Use AI for scale and variation

2. Be transparent

• Tell clients when you use AI tools
• Document your process
• Share which tools you used

3. Edit everything

• Never ship raw AI output
• Refine in traditional design tools
• Add human touches to every piece

The Wrong Way

1. Hiding your process
2. Replacing foundational design elements with AI
3. Ignoring ethical and licensing considerations

the bottom line

The designers who are thriving aren't fighting the last war. They're treating AI as another tool in their kit - something to be learned and applied exceptionally.

What AI image tools are you experimenting with? Hit reply and let me know.

See you next week (and at the Lovable vs. Webflow showdown this Friday),
Tommy

Enjoying this newsletter? Let me know here.​