when developers realize the design was made by ai
a true story about collaboration, confusion, and what modern design really feels like
They opened the link.
Zoomed in.
Paused.
“Wait… what is this?”
That was the moment I knew they realized — the entire Figma file wasn’t drawn by hand.
It was generated.
Prompted.
Conjured into being by an AI designer I’d trained to think like me.
And honestly? I felt a little exposed.
how we got here
The project started like so many do:
An Upwork message.
A fixed-scope contract.
A client named Jimmy M, spiritual guide and founder of The Great Awakener — a media platform blending astrology, channeling, tarot, and global predictions for a massive online audience.
The job sounded simple enough: “Static, developer-ready mockups.”
No prototypes, no animations, no logged-in states — just clean layouts the devs could follow.
That’s how most designers start a freelance engagement.
Except this time, I wasn’t designing the way I used to.
figma make changed everything
Instead of opening Figma and starting from scratch, I opened Figma Make — an AI tool that lets you describe what you need and builds the layout, components, and UI system in seconds.
“Create a Magic Link login page using the existing UI kit.”
“Design a Settings screen with YouTube integration.”
“Generate a video library grid with member-exclusive content blurred out.”
And it did.
Instantly.
Like a design assistant who already understood tone, brand, and behavior.
I didn’t draw The Great Awakener.
I directed it.
Every card, shadow, and glow was part of a larger orchestration — AI generating the skeleton, me refining the soul.
It was fast. Beautiful. Efficient.
Until development began.
when reality caught up
Weeks later, a new project manager named Vaibhav joined the team.
He opened the files. Looked around.
And sent this message in Slack:
“Would it be possible to add clickable prototypes for the main user flows — like adding a comment, viewing a livestream, creating a journal entry, pulling a tarot card, etc.? It’ll help us align across design → dev → QA.”
And that’s when I froze.
Because there were no flows.
No logic trees.
No user journeys.
I’d been hired to make static mockups — and that’s exactly what I’d delivered.
But what the team now needed was a product architecture — something AI can’t fully fabricate.
The irony?
I’d built a system that looked so complete that it gave the illusion of depth.
ai creates coherence, not context
That’s what nobody tells you about AI design.
It’s visually coherent but contextually shallow — a surface that looks like a finished product until you try to click it.
It’s a new kind of uncanny valley:
The pixels make sense, but the logic doesn’t yet exist.
When Vaibhav and the developers started dissecting the files, their questions piled up:
- What happens when a user clicks “Get Started Free”?
- Are predictions clickable?
- What do the dots on the calendar mean?
- If you’re logged in, what does the profile menu show?
- Is the “Unlock full reading” button the same as “Join now”?
All totally valid.
All completely unanswered.
Because the answers didn’t exist yet — they had to be discovered after the design was done.
the problem with working backward
Normally, you start with the experience — the flow — and then design it.
This project flipped that completely.
We started with artifacts and worked backward toward logic.
And that’s where AI throws a wrench into the old process:
It lets you skip ahead visually, but not structurally.
By the time the dev team got involved, the world inside Figma looked alive — dynamic cards, cosmic gradients, blurred member-only content. But it wasn’t an experience yet. It was a vision of one.
So when they tried to build from it, it felt like reconstructing a dream from screenshots.
what it taught me about design in 2025
This project made me realize that we’re in a new design era — one where AI is the paintbrush and the canvas, but humans still write the story.
Figma Make didn’t fail me.
It just accelerated me to the middle of the process — fast.
It let me sketch an entire product ecosystem in days.
But what it couldn’t do was explain how those parts should talk to each other, how the membership logic connects to content, or how the flow should feel when someone actually uses it.
That’s still our job.
The designer’s job.
The human job.
where we go from here
Now, as the project moves into development, I can see both sides clearly.
The developers are trying to build a world that looks like it already exists.
The designer (me) is realizing that beauty without structure creates chaos.
And the AI — silent and efficient — has already moved on to its next prompt.
If I were starting over, I’d do it differently.
I’d begin with a system audit, not a design sprint.
I’d map every journey, screen, and dependency before opening Figma.
And I’d use AI after understanding the architecture — not before.
Because now I know:
AI makes it easy to design something that looks finished, but feels incomplete.
closing thought
When developers realized the design was made by AI, it wasn’t embarrassment I felt.
It was recognition.
We’re all learning how to build in this new era — where prompts replace wireframes, and collaboration means teaching machines to think with us.
The project will evolve.
The team will figure it out.
But I’ll never forget that moment — when someone opened a file and realized they weren’t just looking at a design.
They were looking at the future of how design gets made.
