Artificial intelligence (AI)-powered customer service systems have expanded across Saudi Arabia’s retail sector during 2025–2026, with major retailers increasingly exploring chatbots, recommendation engines, and automated service technologies. Official Saudi digital-economy statistics document the wider shift towards digital commerce, although online-shopping participation should not be interpreted as a direct measure of AI adoption or retailer-specific investment. This study examines the relationship among AI service quality, emotional brand attachment, and brand loyalty using survey results reported for 512 retail customers in Riyadh, Jeddah, and Dammam, collected between May and September 2026. Using partial least squares structural equation modelling (PLS-SEM) and bootstrap resampling, the analysis reports a positive direct association between AI service quality and loyalty (β = 0.421, p < 0.001), together with a positive indirect association through emotional brand attachment (reported indirect effect = 0.238, 95% CI [0.187, 0.289]). The study considers a possible “mechanical efficiency paradox”: functional improvements in AI service may be more strongly associated with loyalty than with emotional attachment. Importantly, this survey does not include a direct comparison with human-delivered service, and the reported positive attachment coefficient (β = 0.187) cannot establish that AI decreases attachment. The reported gender subgroup estimates differ: the AI quality–attachment association is positive for male respondents (β = 0.289, p = 0.001) and statistically nonsignificant for female respondents (β = −0.023, p = 0.811). Age-specific estimates also vary, with more positive reported coefficients in younger groups, although the statistical significance of the age differences requires further supporting output. The manuscript reports that the model accounts for 51.2% of the variance in loyalty (R2 = 0.512) and has positive predictive relevance (Q2 = 0.289). Because some path estimates, correlations, and variance-explained values are not mutually consistent, these results require rechecking against the original SmartPLS output before publication. Managerial implications include offering hybrid AI–human service, allowing customers to choose their preferred channel, and maintaining personal attention for complex or emotionally sensitive interactions. The findings provide a framework for assessing how technological service quality and relationship-building might jointly contribute to brand loyalty.
Artificial intelligence (AI)-powered customer service systems have expanded across Saudi Arabia’s retail sector during 2025–2026, with major retailers increasingly exploring chatbots, recommendation engines, and automated service technologies. Official Saudi digital-economy statistics document the wider shift towards digital commerce, although online-shopping participation should not be interpreted as a direct measure of AI adoption or retailer-specific investment. This study examines the relationship among AI service quality, emotional brand attachment, and brand loyalty using survey results reported for 512 retail customers in Riyadh, Jeddah, and Dammam, collected between May and September 2026. Using partial least squares structural equation modelling (PLS-SEM) and bootstrap resampling, the analysis reports a positive direct association between AI service quality and loyalty (β = 0.421, p < 0.001), together with a positive indirect association through emotional brand attachment (reported indirect effect = 0.238, 95% CI [0.187, 0.289]). The study considers a possible “mechanical efficiency paradox”: functional improvements in AI service may be more strongly associated with loyalty than with emotional attachment. Importantly, this survey does not include a direct comparison with human-delivered service, and the reported positive attachment coefficient (β = 0.187) cannot establish that AI decreases attachment. The reported gender subgroup estimates differ: the AI quality–attachment association is positive for male respondents (β = 0.289, p = 0.001) and statistically nonsignificant for female respondents (β = −0.023, p = 0.811). Age-specific estimates also vary, with more positive reported coefficients in younger groups, although the statistical significance of the age differences requires further supporting output. The manuscript reports that the model accounts for 51.2% of the variance in loyalty (R2 = 0.512) and has positive predictive relevance (Q2 = 0.289). Because some path estimates, correlations, and variance-explained values are not mutually consistent, these results require rechecking against the original SmartPLS output before publication. Managerial implications include offering hybrid AI–human service, allowing customers to choose their preferred channel, and maintaining personal attention for complex or emotionally sensitive interactions. The findings provide a framework for assessing how technological service quality and relationship-building might jointly contribute to brand loyalty.
