APPLICATION OF ARTIFICIAL INTELLIGENCE TOOLS TO IMPROVE SALES FORECASTING ACCURACY IN DIGITAL ENTREPRENEURSHIP

Abstract

The article examines the implementation of artificial intelligence tools for sales forecasting in digital entrepreneurship, focusing on enhancing the accuracy and efficiency of forecasting processes. The research is particularly relevant due to the growing complexity of data analysis in the digital economy and the need for more sophisticated forecasting approaches. The study aims to demonstrate the practical application of AI-based sales forecasting methods in enterprises and to analyze their effectiveness in improving business decision-making processes. The methodology of the research combines theoretical analysis of existing AI forecasting approaches with a practical case study of implementing an automated forecasting system at PJSC «New Style». The study employs machine learning algorithms for data processing and analysis, utilizing both internal enterprise data. The investigation reveals significant improvements in forecasting accuracy and efficiency following the implementation of AI-based tools. The automated system demonstrated a 7% increase in accuracy for product selection planning and a 13% improvement in customer-product pair forecasting. The system also showed a 10% reduction in mean absolute error across markets. Notable results include the complete automation of the forecasting process, reducing manual labor requirements and enabling overnight processing of predictions within 6-12 hours. The system demonstrated remarkable adaptability during crisis periods, successfully recovering its forecasting accuracy within 5-6 months after operational disruptions due to military actions. However, the research also identified limitations in the current implementation, including restricted analysis of external factors, slow adaptation to market changes, and simplified customer segmentation. The practical value of the research lies in its detailed analysis of AI implementation challenges and successes in a real business environment. The findings provide valuable insights for enterprises planning to implement AI-based forecasting systems, offering both technical and strategic recommendations for system deployment. The study contributes to the understanding of AI application in business forecasting and presents a framework for future improvements in automated prediction systems. The research results can be particularly valuable for digital enterprises seeking to enhance their forecasting capabilities and optimize their business processes through AI implementation.