The AI engagement paradox: algorithmic congruence and its inverted-U effect on tourist experience quality

Purpose Despite widespread adoption of AI agents in restaurant and gastronomic tourism settings across Asia, practitioners consistently observe a counterintuitive pattern: increasing AI service involvement initially enhances but subsequently undermines tourist experience quality. This study aims to term this the “AI engagement paradox” and introduce algorithmic hospitality congruence (AHC), a new second-order construct to theorise and empirically confirm the inverted-U relationship between AI congruence and gastronomic tourist experience quality (GTEQ). Design/methodology/approach A sequential exploratory mixed-methods design was used across three Asian markets (Hong Kong, Singapore, Bangkok). Phase 1 comprised 24 semi-structured interviews using the critical incident technique to develop AHC construct items. Phase 2 involved a quantitative survey (N = 412). AHC was grounded in cognitive load theory (CLT). Polynomial regression with response surface methodology confirmed the inverted-U effect; partial least squares structural equation modelling validated the construct. Common-method-bias mitigation and endogeneity-robustness checks used temporal separation, Harman’s single-factor test with a marker variable and Gaussian copula endogeneity correction. Findings AHC comprising culinary AI precision (CAP), emotional service resonance and cultural contextual fit (CCF) demonstrated robust construct and discriminant validity. AHC exhibited a significant inverted-U effect on GTEQ. Cultural collectivism steepened the negative slope, while digital literacy shifted the tipping point rightward. Research limitations/implications This study delivers empirically grounded deployment guidance for hospitality operators in AI-integrated gastronomic settings. The validated 15-item AHC scale provides a diagnostic instrument for monitoring AI-experience alignment. Operators are advised to benchmark their AHC composite score against the estimated tipping-point zone (4.18–5.04 on a seven-point scale), differentiate deployment intensity by tourist digital literacy, apply conservative ceiling margins for high-collectivist segments and treat CCF as a foundational design requirement rather than a localisation afterthought. Digital literacy investment is reframed as a direct hospitality revenue strategy. Originality/value This is among the first studies to theorise and empirically confirm an inverted-U effect of AI congruence in gastronomic tourism. AHC is rigorously differentiated from five related constructs and grounded in CLT from the outset. The findings challenge the linear AI benefit assumption dominant in hospitality AI research and provide indicative deployment guidance for practitioners.