Startups adopting generative AI (GenAI) often face challenges not addressed by existing enterprise-oriented frameworks. Fast feedback cycles, resource constraints, and high uncertainty require a lightweight approach that goes beyond traditional maturity models. This chapter introduces Define-Analyze-Build-Measure-Learn (daBML), an extension of the Build-Measure-Learn (BML) loop, tailored to guide generative AI adoption in startup environments. By adding two upstream phases—Define and Analyze—the framework supports clearer scoping, risk awareness, and value alignment before experimentation begins. The framework is illustrated through two internal use cases from a later-stage software startup, covering domains such as software development, product design, operations, and commercial outreach. These examples demonstrate how daBML can help teams frame GenAI initiatives, identify adoption risks, and track impact through lightweight iterations. While promising in practice, the framework remains conceptual and is not yet empirically validated. Future work should explore how daBML performs across varied startup contexts and how it supports long-term integration of generative AI beyond initial pilots.