Pretty Crap In, Pretty Crap Out
A month ago I was not sure where designers fit in an AI world. So I built 19 projects in 29 days to find out.
I'll be honest. A month ago, I wasn't sure where designers fit in an AI world.
And I do not mean that as a dramatic opening line. I mean I genuinely did not know. I had been reading the same articles everyone else was reading. Watching the same demos. Hearing the same promises about how AI would revolutionize everything. And somewhere in the back of my mind, a quiet voice kept asking, what about us? What about the people who have spent years learning how to listen, observe, and translate human needs into human experiences? Where do we go?
I did not have an answer. And I do not trust people who pretend they do.
So I did what designers do. I prototyped. I tested. I broke things.
I built 19 projects in 29 days, all in GitHub. Just before that, my GitHub account was completely empty. Not to prove a point. Not to post about it. I built them to understand. To feel the edges of what this technology actually is when you stop reading about it and start working with it.
Some were experiments. Quick, rough, ugly things that taught me something and got thrown away. Some were real.
The projects that taught me the most
Some of my biggest highlights came from problems and people I knew deeply.
The first was my mom's website.
That project changed how I think about accessibility entirely. I designed a system where she can send a WhatsApp voice message and, 30 seconds later, it becomes a published article on her website. No login. No interface. Just her voice becoming her words, becoming her content. She said she could finally do it herself. That is the difference between a feature and dignity. No AI would have thought to build that. That came from knowing a person.
The second was the journey management tool.
That came from years of trying to share, manage, and communicate journeys through Miro boards. I knew that problem from the inside. I had lived the complexity of trying to make journeys visible, understandable, and useful across many people. And it was almost impossible to engage non designers in that format. AI did not understand that problem. I did. So I used it to help me design and develop a solution around something I had been trying to solve by hand for years.
The third was helping a colleague with a complex challenge.
I set up a collaborative intelligent system with AI for her so she could experience the technology without limits. The point was never to use AI as a gimmick. The point was to create the right conditions for better thinking, faster movement, and a completely different way of working.
The wall
I also hit a wall.
Early on, I went wild. Building everything. It felt incredible, like having superpowers. But then the costs started adding up. And I had a reality check. Because I can is not a strategy. So I stepped back. I designed my own AI workflow. I got intentional. That constraint made everything better. Because we have always known that constraints are where creativity lives.
What I found
What I found through all of this surprised me.
AI is fast. Genuinely, startlingly fast. Things that would have taken me weeks took hours. The speed is real. The power is real.
But AI does not have soft skills.
It does not have empathy. And I do not mean that in the vague, inspirational poster way. I mean it practically. AI does not understand context the way a human who has spent years listening, observing, and advocating for people does. It does not know that the reason this user hesitates is not a UI problem. It is a trust problem. It does not know that the stakeholder who keeps requesting features is actually afraid of launching. It does not know that the quiet person in the research session has the most important insight in the room.
We know these things. We have trained ourselves to notice them. That is not a nice to have. That is the job.
You still need the brains, the data, the storytelling, and the design. AI just builds what you tell it to. And the quality of what it builds is directly proportional to the quality of what you tell it. If you do not understand the problem, AI will help you build the wrong solution at record speed. If you do understand the problem, AI becomes the most powerful tool you have ever held.
Sharing what we are learning
At the same time, I have also felt extremely inspired by colleagues who have already been working with AI. Seeing what they are doing gave me real hope for all of us collectively. It made me feel that we can solve many problems much faster if we share what we are learning instead of hiding it.
That is why I share these ideas across the company with many different roles. I have shared them with developers, UX designers, service designers, design managers, senior management, Design Operations, and product owners, with the goal of helping us start defining what our roles become in a world shaped by AI.
For me, it is important not to hide what we are learning. This is the moment to share it.
The designers I have been sharing this with, in presentations, in communities, and one on one on their machines, are not excited because AI replaces something. They are excited because it clears space for the work that always mattered most.
The human work.
Pretty crap in, pretty crap out
AI can make bad thinking look polished, but users will still feel the emptiness. That is when you see which companies forgot that users are the business model.
The antidote is us. People who know how to think before they build. People who understand that the hardest part of design was never the pixels.
Nothing is defined yet. The rules have not been written. The role of designer in an AI augmented world is being invented right now by people who show up. We are at the peak of our careers. Not because everything is easy, but because everything is possible.
If you are a designer feeling uncertain right now, you are not behind. You are exactly where you need to be. Uncertainty means you are paying attention.
Now get your hands dirty.