How I Built It: $10K/Month AI Image Generator | Starter Story

Pauline built AI Krea, a service that renovates home interiors in seconds by letting users upload a photo and select a style—Scandinavian, modern, and so on—and see the result within moments. While still a full-time employee at IBM, she grew the app to roughly €8,000 per month in revenue (about $10,000). She recently left IBM to work on it full-time. The conversation covers how she found the idea, validated it, chose pricing and business models, reduced churn, and thinks about opportunity in AI.

Finding and Validating an Idea

Pauline’s approach to finding ideas is grounded in concrete problems rather than abstract inspiration. She looks for problems she or her friends encounter, then asks a series of diagnostic questions: How frequently does this problem occur? Can I build something around it? Can I monetize it? Can I find clients? Only if the answer is yes to all does she start building.

For AI Krea, validation came fast. She built a minimal MVP—simple features, bad quality—and showed it to a few real estate agents she found online or through her personal network. Some said, “That’s awesome. I want to buy it.” And they actually paid. That was the validation she needed. Her principle: “The sooner you ship, the better, because sometimes you imagine your client wants this, but what they actually want is this.” The MVP took about a month; then came a year of improvement.