Development and adoption of AI-driven intelligent agents in tourism: LLM-RAG and PLS-ANN perspectives

Purpose This study investigates the development and adoption of artificial intelligence (AI)-driven intelligent agents in Iraq's hotel sector by integrating Retrieval-Augmented Generation (RAG) and low-code automation to improve operational efficiency, reduce costs, and enhance customer experience. The study further examines behavioural factors influencing adoption using the Unified Theory of Acceptance and Use of Technology (UTAUT) extended with actual usage. Design/methodology/approach A hybrid methodology combining partial least squares structural equation modelling (PLS-SEM) and artificial neural networks (ANN) was employed. Data were collected from 399 hotel managers and decision-makers in Iraq. PLS-SEM assessed the measurement and structural models, while ANNs explored non-linear relationships and determined predictor importance. Findings The findings reveal that performance expectancy, effort expectancy, social influence, and facilitating conditions significantly influence behavioural intention, which strongly predicts actual usage. The ANN results identified social influence as the strongest determinant of adoption. The model demonstrated strong predictive capability, explaining 70.3% of the variance in behavioural intention. Practical implications The study provides actionable insights for managers and policymakers by demonstrating how intelligent agents can automate booking management, email classification, and enquiry handling. Managerial support, training, and robust infrastructure are critical enablers, while cost-effective, low-code AI solutions are particularly suitable for resource-constrained environments such as Iraq. Originality/value This study is among the first to examine AI-driven intelligent agent adoption in Iraq's tourism industry using an integrated PLS-ANN approach, offering theoretical, methodological, and practical contributions for sustainable AI implementation in emerging markets. Peer review The peer review history for this article is available at: Link to the website