Reframing growth hacking in resilient startups: the role of generative AI in experimentation, decision-making and learning

Purpose This paper examines whether and how generative AI (GenAI) enables and reshapes growth hacking in startups aiming to be resilient, focusing on its roles in accelerating experimentation, improving decision quality and orchestrating data, measurement and learning. Design/methodology/approach We adopt an exploratory, multiple-case qualitative design. Data comprise 17 semi-structured interviews with founders and growth leaders across nine startups, complemented by secondary sources. The analysis relies on the Gioia methodology to develop first-order concepts, second-order themes and three aggregate dimensions, which are then mapped onto a seven-stage growth pipeline. Findings GenAI functions as (1) an experimentation accelerator lowering the marginal cost of variation and compressing the idea-to-test cycle, enabling parallel selections of controlled tests; (2) a cognitive sparring partner that reduces bounded rationality and groupthink via premortems, counter-arguments and stakeholder role-plays while preserving human judgment; and (3) a data orchestrator that automates cleaning, cohorting, variance checks and knowledge capture, tightening feedback loops and institutionalizing learning. One important finding is that acceleration in the Generate/Take Action phase translates into durable performance only when Analyze/Prioritize is de-biased by individuals and teams, and Measure/Review converts results into reusable knowledge with appropriate inference discipline. Originality/value The study integrates GenAI uses into a coherent, processual view of growth hacking, showing how AI reallocates human attention from asset production to problem framing, inference quality and organizational learning across the seven stages. We build on decision-making theories that stressed the tension between normative rationality and bounded rationality, to suggest GenAI as a tool to overcome the limited cognitive capacities of individuals and address the overwhelming data volumes.