Artificial Intelligence, Lean Startup Method, and Product Innovations
AI investments complement the lean startup method for innovation, with effects varying by innovation type and AI capability.
Summary
The authors study 1,800 Chinese start-ups between 2011 and 2020, using shifts in government policy that encouraged AI adoption as a natural experiment. Start-ups with strong AI capability produce more innovative products, and AI complements the lean startup method, with the effect depending on the type of AI and the type of innovation. Discovery-oriented AI paired with prototyping reduces uncertainty when exploring new markets, while optimisation-oriented AI paired with controlled experiments such as A/B tests speeds up iterative refinement. The authors conclude that AI is not one thing: firms should match the kind of AI capability to the innovation process it supports, which lets start-ups produce more good products in less time across software and hardware.
Editor's opinion
AI does not raise innovation by itself: it pays off when the type of AI matches the way the firm experiments. AI that searches for new opportunities works with prototyping, and AI that optimises works with controlled experiments. Leaders should choose and organise AI investments around the specific stage of product development they want to accelerate.
Key concepts
AI capability · lean startup method · product innovation · startups