Democratizing market research, without giving up rigor
Panelia was born from a simple observation: quality market research is reserved for those who can pay €10,000 and wait six weeks. We believe testing an idea should be as routine as an A/B test on a web page.
Our mission
Our mission
Turn a product brief into a professional study report (purchase intent, segmentation, verbatims, price sensitivity) in minutes, at near-zero cost, with reliability that's measured and displayed.
Our vision
Make Panelia the default validation layer between an idea and a launch, accessible to everyone, at marginal cost, without sacrificing the science.
The scientific insight
The arXiv 2510.08338 paper shows that properly prompted and calibrated LLMs reproduce human purchase intent distributions. The key isn't to ask the AI for a rating, but to measure the nuance of open-ended responses through semantic similarity.
"LLMs Reproduce Human Purchase Intent via Semantic Similarity Elicitation of Likert Ratings" (arXiv 2510.08338, October 2025).
Authors: B. F. Maier, U. Aslak, L. Fiaschi, N. Rismal, K. Fletcher, C. C. Luhmann, R. Dow, K. Pappas, T. V. Wiecki.
Read the paper (arXiv) ↗An honest scientific posture
We show our limits. Panelia is a decision-support tool, not an oracle; respondents are explicitly synthetic; and we publish our calibration rather than asking for blind trust.
Not an oracle: a decision-support and prioritization tool.
Not a black box: method, limits, and uncertainty are on display.
Not a fake-human generator: personas are explicitly synthetic.
Our 18-month goals
- 01
A multi-segment product with an API
- 02
500 paying accounts and 3 white-label agencies
- 03
A public calibration benchmark vs. human data