Founder Notes
What marketing taught me about building business software
Years of marketing work taught me to distrust demonstrations and trust outcomes. That habit now shapes how I evaluate software, including my own.
Marketing has a way of teaching you to be skeptical of anything that looks impressive but has not been tested against real behaviour. A campaign can look brilliant in a deck and fail completely in market. A landing page can win every internal review and convert worse than the plain one it replaced.
That skepticism followed me into building software. When I look at an AI feature — mine or anyone else's — the first question is rarely “is this impressive.” It is “does this understand the business it is operating in, and does it respect the constraints that business actually has.”
Founders and business owners taught me the second half of that question. Every business I worked with at Digital Dadi had its own version of “we can't do that because…” — a policy, a resource limit, a relationship that mattered more than the metric. Software that ignores those constraints to look automated is not useful. It is a liability wearing a demo.
Building AGE and Humantik, I try to hold both halves of that question at once: understand the business, respect its constraints, and only then ask whether AI can improve the outcome. It is slower than shipping a flashy feature. It is also the only version of this work I trust.