Synthetic Personas for Enhanced Customer Interaction in Innovation: An LLM-Based Approach

This paper investigates the potential of large language models (LLMs) to generate synthetic personas for enhancing customer interaction in innovation processes. Building on existing literature in persona-based design and agent-based simulation, we identify a gap regarding the use of LLMs as interactive customer surrogates. We address this through a prototype system that creates and operationalizes synthetic personas derived from structured templates, enabling dialogue-based exploration of customer needs. An expert interview with an in-dustry practitioner informs the design, while the prototype demonstrates the fea-sibility of leveraging readily available LLM tools, specifically OpenAI models via Azure AI Studio. Results show that generated personas provide contextually appropriate, differentiated feedback aligned with domain knowledge, suggesting their value as a complement—not a replacement—to direct customer involve-ment. We discuss integration into the enhanced FuGle innovation model and highlight ethical, trust, and validation challenges. The study establishes a foun-dation for further research on LLM-driven customer proxies in innovation man-agement.