Launching Under AI Uncertainty

Sammanfattning

This study explores how software startup teams make sense of AI-related uncertainty in the pre-launch phase. While existing research often presents AI as a tool for reducing uncertainty through improved information processing, prediction, and efficiency, the findings of this study show that AI simultaneously creates new forms of ambiguity around trust, competitive differentiation, product reliability, and strategic relevance. Drawing on sensemaking theory, the study examines how startup teams interpret AI-related cues, negotiate competing interpretations, and construct understandings that become plausible enough to support action under conditions of rapid technological change.The study adopts an exploratory qualitative multiple-case design based on 15 semi-structured interviews across 11 software startups, complemented by full-day observations in selected cases. The findings are organized around three sequential empirical themes. First, three primary cues triggered AI adoption among startup teams: fear of missing out, efficiency gains and lower time to market, and competency gaps. Second, through trial and error, startup teams developed conditional trust in AI and integrated it as an active participant in everyday work, shaping idea generation, coding, prioritization, evaluation, and decision support, while consistently maintaining human verification practices for strategic and customer-facing decisions. Third, plausibility was negotiated through human screening, provisional acceleration, and time pressure, with startup teams acting on what was sufficiently actionable rather than waiting for uncertainty to be fully resolved.The central contribution of this study is that AI redistributes instead of resolving uncertainty in the pre-launch phase. Across all eleven firms, a consistent and recurring pattern emerged: AI reduced uncertainty at the micro operational level, where tasks were bounded, testable, and correctable, while simultaneously increasing uncertainty at the macro strategic level, where concerns about market instability, competitive imitation, and the durability of competitive advantage could not be tested or resolved in the same way. Startup teams did not resolve this asymmetry but continuously navigated it, constructing operational plausibility while carrying unresolved AI-related strategic uncertainty forward as a background condition of the pre-launch phase.