Work

From tool to co-designer: generative AI in early-stage startup innovation

Year: 2026
Type: dissertation
Abstract: This thesis examines how early-stage startups integrate generative artificial intelligence (AI) into innovation practice and how this integration reshapes cognition, professional roles, and collaborative dynamics. While prior research has often treated AI primarily as a tool for efficiency, automation, or analytics, this study approaches generative AI as a co-designer that increasingly participates in ideation, reframing, prototyping, and strategic sensemaking within startup environments. The study addresses a gap in existing literature by focusing on the everyday, practice-based use of AI in resource-constrained and rapidly evolving startup contexts. The research adopts a qualitative, exploratory, and abductive approach. Empirical data consists of sixteen semi-structured interviews with founders, product managers, designers, and other startup professionals from different industries. The data was analyzed using the Gioia methodology in order to identify first-order concepts, second-order themes, and aggregate dimensions related to human-AI collaboration in innovation work. The findings show that generative AI is no longer used merely as an operational tool, but increasingly as a cognitive partner that supports externalization, reframing, and the expansion of the ideation space. AI accelerates iteration and prototyping, compresses innovation cycles, and alters the architecture of innovation from more linear stage-based processes toward continuous and non-linear learning loops. At the same time, the integration of AI transforms human roles: startup professionals move from being primary creators toward acting as curators, interpreters, and evaluators of AI- generated alternatives. Human judgment becomes more critical, not less. However, the study also identifies significant tensions. These include cognitive overload, decision fatigue, trust-related problems, the illusion of correctness, authorship ambiguity, and the risk of homogenized or convergent ideas. These findings therefore suggest that AI expands creative exploration while simultaneously increasing the need for critical filtering, contextual interpretation, and ethical judgment. Based on these insights, the thesis proposes a conceptual framework for human-AI co-creation in early-stage startup innovation, highlighting the importance of stage-specific role distribution, iterative co-creation loops, and adaptive safeguards for responsible and strategically effective AI use. This thesis contributes to research on AI, design, and entrepreneurship by extending the discussion of AI as a co-designer into the context of early-stage startups. Practically, it offers guidance for startups seeking to integrate generative AI in ways that enhance innovation without undermining human agency, originality, or reflective thinking.
Institution: Aalto University
Language: English
Open Access status: green
OpenAlex ID: W7161364111
Cites: 78
Cited by:
Related to: 10