Business Amplification: Managing Innovation through Synthetic Users

This paper examines the potential of Synthetic Users (SyU) as a possible alternative or supplement to conventional human surveys in the area of innovative product or service development. Generative Artificial Intelligence (GenAI) can be an invaluable technology for companies and creative businesses during the “Eureka phase” and in testing, economic analysis, development, and manufacturing. GenAI can rapidly answer hundreds of questions without causing fatigue, creating rapid feedback and making it a cost-effective alternative to human testing (Aher et al., 2023). The deployment of SyU promises a more efficient business world, offering benefits such as predictive and cultural insights, and faster iteration, contributing to a deeper understanding of the emotional drivers behind user behaviours. SyU could be an innovative approach that blurs the boundaries between qualitative and quantitative investigation by allowing researchers to gain rich insights and leverage large-scale data, lowering biases in human interactions, observer-expectancy effects, emphasizing the importance of diverse data, objective questioning, and regular evaluation (Santurkar, S. et al., 2023). Finally, while conducting a survey again under identical conditions in the real world is not always feasible, in the synthetic realm it is possible to improve the LLMs approach through a recent technique known as "self refinement" (selfrefine.info), based on an iterative algorithm, punctuated by two moments, feedback and refinement, that result in higher quality output and a more structured process. (Maadan et al., 2023) After reviewing the current state of the art, the article explores the potential of Large Language models (LLMs) to simulate human-like interactions and judgments through a real case. One company, Syntheticusers.com, based on GPT-4, has been selected to test the comparison between human participants and SyU in evaluating an immersive museum experience within a 5G-focused project led by the University of Turin. Our results suggest that scientists study the right mix of humans and artificial intelligence, rather than focusing on whether artificial intelligence will replace humans.