Generative AI in digital engagement: a quasi-experimental study of tourist sentiment

Ali, W. ORCID: https://orcid.org/0000-0002-6200-8580, Kasturiratne, D. ORCID: https://orcid.org/0009-0000-1959-0918, Ameer, I. ORCID: https://orcid.org/0000-0002-2379-7863 and Bhaskar, S. ORCID: https://orcid.org/0000-0002-8835-1918 (2026) Generative AI in digital engagement: a quasi-experimental study of tourist sentiment. The Service Industries Journal. ISSN 1743-9507 doi: 10.1080/02642069.2026.2638226

Abstract/Summary

This study examines whether generative AI can enhance tourist sentiment in online reviews by acting as a consistent and scalable form of digital engagement. Using a quasi-experimental Difference-in-Differences design, we analysed 11,393 reviews collected for six months from five Indian restaurants in a UK tourist city, where one restaurant adopted ChatGPT-generated responses and four served as controls. Results show that AI mediated responses produced a significant improvement in tourist sentiment, with similar effects across both social (Google, Facebook) and delivery-oriented platforms (Deliveroo, Just Eat, Uber Eats). Response features such as tone, personalisation, and length had limited additional influence, indicating that the presence of a response matters more than its specific stylistic qualities. The findings suggest that GAI-mediated responses influence tourist sentiment primarily by signalling organisational attentiveness and relational legitimacy, rather than through nuanced stylistic features of the response. The study demonstrates how AI can support post-visit engagement in tourism settings and offers practical guidance for firms seeking efficient strategies to manage online reviews and strengthen their digital service presence.

Item Type	Article
URI	https://centaur.reading.ac.uk/id/eprint/128929
Identification Number/DOI	10.1080/02642069.2026.2638226
Refereed	Yes
Divisions	Life Sciences > School of Agriculture, Policy and Development > Department of Agri-Food Economics & Marketing
Publisher	Routledge
Download/View statistics	View download statistics for this item