Academic ResearchAugust 7, 2026arXiv (cs.IR, cs.CY)

Invisible to the Machine: Auditing AI Restaurant, Cafe, and Bar Recommendation Against a Complete Market Census

The first comprehensive audit of AI venue recommendations against a complete market census: 4,776 food and beverage establishments across two Bali locations, tested through 96 queries to ChatGPT, Claude, Gemini, and Perplexity. 85.6% of venues were never recommended by any system, visibility hinged on review counts, website presence, and price signals rather than star ratings, and systems sometimes surfaced permanently closed venues with low cross-system agreement.

Authors

Vladimir Pitenin

Article content

This paper presents the first comprehensive audit of AI venue-recommendation systems for restaurants, cafes, and bars, benchmarked against a complete market census. The study examines 4,776 food and beverage establishments across two locations in Bali and tests responses from ChatGPT, Claude, Gemini, and Perplexity to 96 diverse queries. A striking 85.6% of venues were never recommended by any system.

Visibility in AI outputs depends heavily on review counts, website presence, and available price information, while star ratings show minimal effect on whether a venue appears in initial recommendations. The systems occasionally suggested permanently closed establishments, and cross-system agreement on recommended venues remained low.

The results expose a structural visibility bias in generative-AI local discovery: operators without a strong review footprint, a working website, or transparent pricing risk becoming invisible to the machine layer that increasingly mediates guest choice, with clear implications for hospitality marketing, discoverability strategy, and platform accountability.

A Systematic literature review of AI-based service failure and service recovery in hospitality and tourism industries using the TCCM framework

A PRISMA-based systematic review of 78 empirical studies (2018–2025) on AI-based service failure and recovery in hospitality and tourism, synthesized through the TCCM (Theory, Context, Characteristics, Methodology) framework. The review shows the field is fragmented across contexts, constructs, and methods, and dominated by scenario-based experiments — and it lays out a TCCM-informed research agenda plus practical guidance for scenario-specific AI service recovery strategies.

Academic ResearchJune 6, 2026

When Robots Speak: Exploring Conversational Styles, Robot Persona Realism, and Engagement in Hospitality and Tourism Services

Across one field study and three online experiments in restaurants, hotels, railways, and airports, the authors show that emotional conversational styles in AI service robots drive higher purchase intentions than rational styles, with engagement as the mediating mechanism. Robot persona realism moderates the effect: emotional styles work best with highly humanlike robots, while rational styles outperform with less anthropomorphic ones — a style-persona congruence rule for designing conversational AI in hospitality.

For Professors

Submit article for consideration

If you are a professor or researcher and would like to suggest a publicly available article for inclusion in the Research Hub, you can submit it for review and possible inclusion through our dedicated submission form.

Submit research