The first place AI should live isn’t in your product.

Almost every founder I speak to asks the same question, “How should we implement AI?”

Most founders think the answer lives in the feature set. It doesn’t. It lives in how we make decisions. When we first started experimenting, we did what everyone does. We generated mockups faster, used AI to write content, played with Figma Make and tested tools like Lovable.

It felt productive. It gave us flows in minutes, clean screens and confident demos. It felt like progress. AI gave us velocity and gave stakeholders the illusion of efficiency.

With AI, you can generate a huge number of polished directions in an hour. But we weren’t moving faster, we were only focusing on what we could see and losing the holistic view. We spent more time going back and fixing things than we would have spent without AI at all.

That’s when it clicked.

AI wasn’t exposing a design problem, it was exposing ambiguity. When you lose sight of the overall picture and just fix isolated problems, you multiply noise.

That was the shift. We stopped asking, “Where can we add AI?” And we started asking, “Where does clarity break down?” It wasn’t in polish, it wasn’t in the number of screens we could produce, it was in eliminating ambiguity and stress-testing assumptions.

This is where we saw results. Clearer communication across product and engineering. Faster alignment in decision meetings. Stronger thinking. Automated assets and documentation for our design system that reduced complexity instead of increasing it.

Knowing where Figma Make sits in our process. Knowing how MCP servers fit into our architecture. Knowing how to leverage tools like Cursor with engineers. Knowing when automation supports thinking and when it replaces it.

That's the real insight.

Once we understood this, we moved away from speculative output and closer toward production clarity. The change to our process was subtle, but the impact was massive. AI became a thinking accelerator, not a magic wand.

So when you look at how to use AI, the first place it should live is not in your product, it's in your decision making. If it doesn’t, you’ll move faster and fall behind at the same time.