case studies
Comparing human and synthetic service responses
Five empirical service studies provide a benchmark for exploring where synthetic agents reproduce human responses and where they diverge.
Can synthetic agents help researchers study service failures and recoveries?
The study
Researchers used SurveyLM to generate synthetic responses to scenario-based experiments from five empirical studies, comparing the results with the original manipulations and hypotheses.
What it found
Alignment varied across contexts. Some text-based scenarios produced closer agreement, while scenarios depending on visual or physical context exposed limitations. Plausible responses alone did not establish fidelity to human behaviour.
Why it matters
The study positions synthetic agents as a research instrument requiring empirical comparison, careful design and sensitivity analysis. It does not establish a universal replacement for human participants.
Publication
Comparing human and synthetic data in service research: Using augmented language models to study service failures and recoveries.
Steven J. Bickley, Ho Fai Chan, Bang Dao, Benno Torgler, Son Tran and Alexandra Zimbatu. Journal of Services Marketing, 39(1), 36–52 (2025).