Abstract
<jats:p>This article addresses the topical issue of adaptive context-aware recommendation in CRM systems, specifically examining the challenges of using context-aware computing in CRM systems, including in combination with artificial intelligence technologies. An adaptive method for context-aware recommendation of customer interaction strategies in CRM systems is proposed; specifically, a mechanism for recommending interaction strategies based on metrics and indicators of customer activity and taking into account customers’ belong to different groups has been developed. The recommendations are made context-aware, which ensures greater accuracy of recommendations and deeper personalization in working with customers in CRM systems. A method has been developed to adapt the computational load to the service priorities of different customer groups and the current system load in order to reduce the computational resource consumption of the context-aware recommendation method. The possibility and prospects of applying machine learning methods at various stages of the proposed context-aware recommendation method have been considered. A prototype CRM system has been developed in which the proposed context-aware recommendation method has been implemented in software. An experimental study of the developed adaptive context-aware recommendation was conducted, which showed that it outperforms the non-context-dependent counterpart by 3.4% in conversion rate and by 20% in positive predictive value (PPV), and also reduces the computational load by an average of 11% compared to the non-adaptive mode. Key words: context-aware computing, customer relationship management, machine learning, recommender system</jats:p>