No PM. No eng team. No one to tell me my ideas are bad. So I built a system.
It’s been a year since I graduated and I haven’t secured a design role yet. No experience means no PM across from me to pressure test ideas, no engineering lead to tell me what's actually shippable, no design director catching my blind spots.
What I do have are ideas, an urge to get better, a few side projects and a workflow I kind of stumbled into that's genuinely changed how I think.
The problem with building alone
When you're solo, the hard part isn't the work. It's the thinking. There's no one to ask “wait, why are we doing it this way?” No one to hit the brakes when you're in love with an idea that doesn't hold up. You can go deep on something for days and realize you were solving the wrong problem the whole time.
I started using Claude the way most people do: as a generator. Spit out ideas, pick the best one, keep moving. It felt efficient. I also produced a lot of mediocre work.
So I switched up my thinking.
Decide → Challenge → Iterate
The shift was small but it changed everything. I stopped asking Claude to generate ideas first and started bringing my own half-formed decisions to it instead.
Here's how it actually works:
I show up with a direction. I explain my reasoning. Then I ask Claude to push back, not to validate me, but to find the holes. I literally tell it: “Be as critical as you can.”
Sometimes the pushback confirms I'm on the right track. More often it shows me something I missed: an edge case, an assumption I'd baked in without realizing, a constraint I glossed over.
Then I iterate. Not because Claude told me to, but because I actually understand my own reasoning better now.
The output isn't just a tighter feature spec. It's a sharper mental model.
What this is (and isn't)
Yes, I know this is not the same as having a PM.
A real PM brings way more: organizational context, relationship capital, actual stakes. When they push back, it costs something. When I push back on myself through Claude, I can just ignore it. That gap is real and worth acknowledging.
But what this is, is a forcing function for solo thinking. A way to stress test ideas without waiting for a crit, a meeting, or someone else's calendar to free up.
For someone early in their career, building without a team, trying to develop product intuition without the reps that come from a real role? That's genuinely useful.
The one thing that makes it work
The prompt matters. I don't ask “what do you think?” I say:
Here's my decision and my reasoning. What's the strongest case against it? What am I not seeing?
That's it. The specificity of the challenge is what changes the quality of the answer.
The goal isn't to outsource my thinking. It's to make it harder to poke holes in.
Building alone is hard. Thinking alone is harder. This is just one way I've made it a little less lonely and a lot more rigorous.