Open-source platform uses genAI to automate conjoint survey design

Conjoint analysis—a staple in marketing, political science, and HCI—has long been hampered by expensive commercial platforms and complex survey infrastructure. Philipp Brauner's new open-source platform changes that by combining generative AI with a self-hosted web application. Researchers can define a base prompt, which the system parameterizes with conjoint profiles to produce both textual scenario descriptions (via an LLM) and visual stimuli (via a text-to-image model). Optional LLF-facing level annotations further refine the outputs.

Beyond stimuli generation, the platform offers a structured setup wizard, AI-assisted attribute suggestions, and live data analysis dashboards—all designed to lower the barrier for novices. A proof-of-concept study on care robot preferences for ambient assisted living (N=55) demonstrates the system's viability. The author emphasizes that while AI augments stimulus creation, theoretical grounding remains the researcher's responsibility. Full export bundles (stimuli, prompts, response data) ensure transparency and reproducibility, opening new methodological avenues for HCI and related fields.