Koah discovered that a basic RAG system for ad serving is ineffective. Embedding models can match keywords (e.g., 'Civil War' query to a 'Civil War' game) but fail to grasp the user's underlying intent (e.g., homework vs. entertainment). True contextual relevance requires more sophisticated AI.
Koah's founders developed deep empathy for their customers by first failing to monetize their own apps. This firsthand struggle with unsustainable "dark patterns" gave them the conviction and nuanced understanding necessary to build a product that truly solves the problem for other developers.
To combat the startup tendency of building too many features, Koah's CPO forces the team to answer, "If you could literally only work on one thing, what would it be?" This constraint cuts through noise and exposes the true top priority, accelerating focused development.
Unlike transactional search engine queries, user interactions with AI tools are typically top-of-funnel and exploratory. Monetization strategies must acknowledge this by guiding users from curiosity to purchase, rather than expecting direct conversions as seen in traditional search advertising.
Instead of viewing competitors as threats, Koah sees them as beneficial. In a new space like AI advertising, rival companies help educate the market and normalize the business model. Their collective efforts build legitimacy faster than a single company could alone.
Koah's CPO suggests replacing the term "ads" with "sponsored commercial experiences." This reframes monetization from intrusive banners to interactive, value-adding engagements, making it more palatable for developers and advertisers in the new AI paradigm.
Koah's ability to build and iterate quickly isn't an accident; it's a "muscle" developed from their prior experience of launching a new app every month. This history of constant, low-stakes launches normalized failure and overcame the fear of shipping imperfect products.