Leveraging large language models for cross-cultural innovation research: a cost-effective approach to understanding market perceptions

Purpose This study assesses how Large Language Models (LLMs) can support cross-cultural market sensing by approximating social survey responses. Design/methodology/approach Responses from five LLMs (ChatGPT, Gemini, Claude, LLaMA, DeepSeek) were compared with the European Social Survey (ESS) Round 11 data for nine questions on gender perceptions in healthcare, employment, and law enforcement in Spain and Finland. Statistical similarity was assessed using the Jensen–Shannon Distance (JSD), Hellinger Distance (HD), and Relative Mean Absolute Difference (RMAD). Findings LLM outputs can approximate survey distributions but cannot substitute for them; in 17 of 18 question–country pairs, at least one model attained high or very high similarity (JSD and HD < 0.2; RMAD <5%). This represents the upper envelope of the five-model portfolio rather than the performance of any single model selected. Models replicated personal experiences better than collective cultural perceptions. No model consistently outperformed across contexts, and substantial divergence across model-question combinations made this instability a central finding. Research limitations/implications This study covers two countries and one thematic module. Future research could examine other contexts and domains. Practical implications LLMs can support lower-cost preliminary market sensing, with validation across multiple models. Social implications LLMs can broaden access to faster, lower-cost cross-cultural research but require human oversight and validation. Originality/value This study provides a systematic multi-model benchmark against probability-sampled cross-national survey data. Its contribution is empirical and methodological, with an exploratory dynamic capability interpretation and an application of generative AI to organisational market sensing.

Authors

Institutions

Publication Details

Journal
European Journal of Innovation Management
Published
2026-11-20
DOI
https://doi.org/10.1108/ejim-06-2026-0854
Primary Topic
Topic Modeling
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Leveraging large language models for cross-cultural innovation research: a cost-effective approach to understanding market perceptions

Kathrin Komp, Macarena Estévez, Jorge Sáinz, María Teresa Ballestar
European Journal of Innovation Management
Topic Modeling
article

Leveraging large language models for cross-cultural innovation research: a cost-effective approach to understanding market perceptions

Kathrin Komp, Macarena Estévez, Jorge Sáinz, María Teresa Ballestar
article en

Abstract

Purpose This study assesses how Large Language Models (LLMs) can support cross-cultural market sensing by approximating social survey responses. Design/methodology/approach Responses from five LLMs (ChatGPT, Gemini, Claude, LLaMA, DeepSeek) were compared with the European Social Survey (ESS) Round 11 data for nine questions on gender perceptions in healthcare, employment, and law enforcement in Spain and Finland. Statistical similarity was assessed using the Jensen–Shannon Distance (JSD), Hellinger Distance (HD), and Relative Mean Absolute Difference (RMAD). Findings LLM outputs can approximate survey distributions but cannot substitute for them; in 17 of 18 question–country pairs, at least one model attained high or very high similarity (JSD and HD < 0.2; RMAD <5%). This represents the upper envelope of the five-model portfolio rather than the performance of any single model selected. Models replicated personal experiences better than collective cultural perceptions. No model consistently outperformed across contexts, and substantial divergence across model-question combinations made this instability a central finding. Research limitations/implications This study covers two countries and one thematic module. Future research could examine other contexts and domains. Practical implications LLMs can support lower-cost preliminary market sensing, with validation across multiple models. Social implications LLMs can broaden access to faster, lower-cost cross-cultural research but require human oversight and validation. Originality/value This study provides a systematic multi-model benchmark against probability-sampled cross-national survey data. Its contribution is empirical and methodological, with an exploratory dynamic capability interpretation and an application of generative AI to organisational market sensing.

European Journal of Innovation Management
Universidad Rey Juan Carlos (ES), Lappeenranta-Lahti University of Technology (FI)
Topic Modeling
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon