Software as Clay
Notes on the builder mindset in the age of AI
I’ve been thinking a lot lately about what it means to have a builder mindset.
As a player-coach inside Pendo’s design team, I’ve been watching my team adapt. Designers are getting closer to code. Product people are moving from customer problem to prototype faster than ever. Engineers are tightening the loop between shipping, testing, and learning.
I’ve been adapting too.
As a manager, I still coach, align, give feedback, and create clarity. But I also find myself moving across the PM / UX / content / front-end engineering spectrum: shaping ideas, prototyping, writing, testing workflows, and practicing my way into this new way of building.
Because AI is changing what software feels like to work with: less like a fixed object we carefully carve, and more like a responsive material we can shape, test, and reshape as we learn.
For years, we treated software like marble.
We planned the cuts. We measured twice. We wrapped the work in documentation, rituals, reviews, and roadmap debates because each decision felt costly and difficult to undo.
And for a long time, that made sense.
Software was expensive. Engineering capacity was scarce. Mistakes were costly. So we slowed down and tried to reduce uncertainty before touching the material.
But AI is turning software into something more like clay.
Clay can be shaped, tested, collapsed, reworked, and shaped again. We can subtract. We can add. We can change direction as the form starts to reveal itself.
Clay rewards contact.
It rewards motion.
It rewards learning through making.
And when the material changes, the way we work has to change too.
The central question is shifting from:
“Can we build this?”
to:
Do we understand the customer, the workflow, the business, and the problem deeply enough to know what is worth building?
That understanding is the work now.
AI can help generate code, interfaces, summaries, prototypes, launch-copy, research synthesis, and workflow ideas. But it can’t replace the context that comes from being close to the customer, close to the workflow, close to the business, and close to the consequences of what you ship.
The people with that context can’t afford to stay passive.
We can’t wait for perfection — the perfect handoff, the perfect ticket, the perfect brief, or the perfect moment when the solution becomes verified and clear.
We have to get closer to the clay.
That means owning more than our assigned tasks.
If you own a feature, a workflow, a product surface, or a customer problem, you can’t only ask:
“What ticket am I supposed to complete?”
You have to ask:
How do I make this experience unmistakably better?
Where are users getting stuck?
What would I change today if my livelihood depended on the success of this workflow?
What do I know that no one else is acting on yet?
What is the smallest thing I can ship, test, or learn from right now?
What would I change if I were responsible for whether this actually works?
That is the builder mindset.
It’s the difference between doing the assigned work and caring about what happens after it ships.
And yes, that can feel exposed.
The old process gives us cover. So do our roles. A role tells us where we stand, where our authority begins and ends, and how we’re supposed to show up. That clarity is useful. It helps teams coordinate. It gives us stability.
But it can also become a canopy.
Underneath it, we can avoid being too wrong, too visible, too early, or too exposed. We can point to the plan, the process, or the edges of our role when we feel unsure.
Real building asks us to reach beyond that canopy.
Not to abandon expertise. Not to dissolve every boundary. But to move beyond the narrowest interpretation of our job when the work calls for it.
AI does not make building easy.
In some ways, it makes building harder. The blank page gets bigger. The number of possible directions explodes. The burden of judgment increases.
But it also gives us something back.
Motion.
Experimentation.
Direct contact with the material.
So maybe the invitation is simple:
Stop approaching software like marble.
Bring your context, judgment, customer understanding, and agency to the table.
Then start molding the clay.