How to land your first customers before the product exists

The spark for Pathfindr was a client's 40-year-old Pascal codebase fed into ChatGPT to see what would happen. The idea itself was sketched on a notepad in a coffee shop: clients on one side of the page, AI on the other. Dawid Naude gave his first keynote talk with nothing built yet - just the sketch and the pitch. By the lunch break that followed it, he had three paying customers.

Talk where your customers already are

Dawid didn't wait for an audience to find him - he went and stood in front of the room his future customers were already sitting in. A keynote is three things at once: a forcing function to clarify the idea, a credibility signal, and a captive room of exactly the people who feel the problem you're describing. One talk did the work a cold outreach campaign usually spreads across months.

Sell the outcome, not the software

Nobody bought software in that lunch break, because there wasn't any. They bought the outcome on the right-hand side of the notepad - what their business would look like once AI was doing the work their old processes couldn't scale. When the product is a sketch, the sale has to rest entirely on the problem being real and the person selling it being believable. If you can't close on that, more features won't save the pitch either.

Speed to first revenue beats polish

Three paying customers before a line of code is the strongest validation a founder can buy, because money is the one signal that can't be polite. It also shapes what gets built: the first version of Pathfindr was built for people already paying for it, not for an imagined persona. Eighteen months later Affinda Group acquired the company in an all-scrip deal worth $15 million, and Dawid rolled his equity straight back into the combined group. The sequence matters - customers first, product second, exit third - and every step of it started with standing in a room and talking.