IMPLEMENTATION AND EVALUATION OF THE EFFECTIVENESS OF AN ADAPTIVE BPM SYSTEM FOR SMEs: A CASE STUDY BASED ON ML AND SERVERLESS WITH CYBERSECURITY IN MIND

The article presents a case study on the implementation and evaluation of the effectiveness of an adaptive business process management system (BPM) for small and medium-sized businesses (SMEs), integrating machine learning (ML), serverless architecture and cybersecurity measures. The research is based on the previous works of the author [1, 2], where the methods of designing automated information systems (AIS) were analyzed and a conceptual model of the system was developed. The relevance of the topic is due to the growing need for automation in SMEs in developing economies such as Kazakhstan and Russia, where traditional ERP/CRM systems are often ineffective due to high costs and low adaptability. The goal is to evaluate the ROI of the system through implementation in two fictional but realistic cases (retail and services) using data modeling for privacy.The methodology includes the collection of metrics before and after implementation (process processing time, costs, security level), A/B testing, and log analysis. Python with TensorFlow for predictive analysis ML models, AWS Lambda for serverless components, and OWASP threat assessment tools were used. The results show a 25-30% reduction in costs, a 35-40% improvement in adaptability due to ML, and an increase in cybersecurity through vulnerability detection. Ethical aspects include data anonymization and compliance with GDPR-like standards. The findings confirm the benefits of the system for SMEs, with recommendations for scaling and discussion of limitations such as reliance on cloud services.