Design and Validation of an Artificial Intelligence-Based Customer Relationship Management Model for Competitiveness in the Food Industry

Abstract

The present study was conducted with the aim of designing and validating a customer relationship management (CRM) model based on artificial intelligence (AI) to enhance the competitiveness of the food industry. This study employed a mixed-methods approach (a combination of qualitative and quantitative methods). In the qualitative phase, the research was fundamental in nature, while in the quantitative phase, it was applied in its objective. In the qualitative section, grounded theory methodology was used, focusing on in-depth and semi-structured interviews. In the quantitative phase, structural equation modeling (SEM) was employed to design the AI-based CRM model aimed at achieving competitive advantage in the food industry. Overall, the study was carried out in two main phases. In the first phase, through a systematic review of theoretical foundations and prior research, key concepts and preliminary thematic areas related to the subject were extracted. Subsequently, in-depth and semi-structured interviews and expert consultations were conducted, and the collected interviews were subjected to thorough analysis. The statistical population for this research included experts in the food industry sector, all managers and specialists at the Ministry of Health and Medical Education (Food and Drug Administration), managers and experts from the Cultural Heritage, Handicrafts and Tourism Organization, and private sector actors active in the food industry. Sampling was conducted using the snowball method. The researcher achieved theoretical saturation after 13 interviews.  In the quantitative phase, the statistical population included CRM experts in the food industry, from which a random sample of 385 individuals was selected. For data analysis in the qualitative phase, a systematic method was used through three main coding steps: open coding, axial coding, and selective coding. In the quantitative phase, the proposed model was tested using the structural equation modeling approach via SmartPLS software. The results of the qualitative phase revealed that, during the open coding stage, more than 49 subcategories were identified. During axial coding, a more precise examination of the identified codes was performed, and by aggregating and categorizing the related open codes, ten core categories were established. Furthermore, the findings of the quantitative phase indicated that the proposed research model had a good fit.