Artificial Intelligence, Lean Method and Startup Product Innovation

Xiaoning (Gavin) Wang, PhD Candidate, and Lynn Wu, Operations, Information and Decisions, The Wharton School

Abstract: New product development can be defined as the set of activities that start with responding to a market opportunity and end with the delivery of a differentiated product or service. Such activities are crucial for startups not only because of the huge market gains that a successful product could provide, but also that a failed product may directly lead to the depletion of cash reserves and bankruptcy. With the accelerating pace of technological advancements, an increasing number of startups have adopted artificial intelligence (AI) to develop products in various industrial settings, including autonomous vehicles, facial recognitions and natural language processing, etc. However, not everyone has witnessed business gains from their AI investments, and recent work also documents the evidence of systematic differences in the early return of AI adoption. Such observations reflect a research gap in the academic literature to explain the variations in company-level AI returns due to organizational factors. Moreover, most of the existing empirical research restricts their objectives to public companies, while few paper has examined the effect of AI adoption in startups. In this study we close the research gap by examining the effect of AI adoption in the startup product development process, and how such effect is moderated by a commonly adopted business practice in early-stage firms: lean method. Originated from the software startups, lean method is an experimentation-based approach to rapidly respond to customer demands in product development. Lean methods are shown to be especially effective when market uncertainty is high. Similar to lean methods, using AI to automate predictions can help firms learn about market dynamics by tapping on the pulse of consumer and economic behaviors. As startups are at the forefront of adopting AI tools, it is important to understand whether AI can amplify the effects of lean methods in helping startups to be even more responsive to market conditions that before. And if so, to what extent and under what conditions AI can complement lean method in developing products. The relationship between AI and lean method could partially explain the uneven returns of AI across businesses as startups need to decide whether to concentrate resources on one of them or to invest in both as a system.