Last week, I wrote about building my first real app / bigger project with AI tools. That project gave me the momentum I needed to finally build the thing I had committed to back in December – a real enterprise website.

I built it. And I got it completely wrong.

I chose this beautiful blue gradient and spent almost an hour blending it into the hero section. It felt subtle, intentional – simple, in my mind. In reality, my team prefers things actually simple. No gradients, no extra flourish. Just clarity. I thought I was aligning with that. I wasn’t.

At first, it still felt like a win. I had finally figured out how to connect the backend to the frontend, which unlocked something for me. I moved quickly and leaned into my own taste – colors, fonts, layout. I had a point of view, and in a world where AI can easily produce something generic, I wanted this to feel human.

But it was only my perspective.

I work on a team with shared ownership over what we put into the world. And while I told myself everything was open to feedback, I had already run too far ahead. When I presented it, the reaction wasn’t alignment – it was confusion. They thought I was building a backend with more room to experiment. They didn’t realize I had redesigned the entire homepage. What can I say – I’ve always been an extra credit kinda gal.

Looking back, the disconnect makes sense. I had been told not to go off and build this entirely on my own. But I did it anyway because I wanted momentum. I wanted something tangible. I wanted to see progress. What I didn’t account for is how much AI accelerates that instinct. I had built an MVP before we even aligned on a scope of work.

Being a few steps ahead is fun. But when you’re too far ahead, you’re wearing a full winter coat – gloves, scarf, everything – and everyone else is still in summer mode, getting a tan in their trunks. You have to be aligned on timing. Otherwise, you don’t just move fast – you move alone.

That’s the mistake I made, and it clarified something important.

There are two very different ways of building.

The first is the “dream” space: where you’re building something aspirational, expressive, a version of what you or your company could become. When I built the site, I chose fonts, colors, and layouts based on what I believed we are today and who we could be over the next few years. But I didn’t bring in the people who have been building this company for over a decade. They felt left out – and they were right. A font can seem trivial, but to them it was an intentional decision made years ago. Understanding that is part of the process. The process I skipped.

The second is the “optimization” space: where and when you’re solving for efficiency, it’s for a very specific problem. For example, I built a small tool this week to prep my team before meetings with founders. It pulls together context, summarizes key points, and turns it into something usable before the call. It’s tightly scoped and very specific. In that space, friction is less useful. Speed and precision matter more.

Both modes are valuable, but they require different approaches. What went wrong with my website is that I treated a dream-space project like an optimization task. I optimized for speed when I should have optimized for alignment.

Before you start building, it’s worth asking which mode you’re in. Are you exploring something new, or solving something specific? Some projects need more friction. Some need less. Knowing the difference is what keeps you from getting all the way to the end – only to realize you built the wrong thing.