Deep Learning Model for Enterprise Process Performance Prediction: Graph Convolutional Network Approach

Enterprise prediction is based on the use of smart analytics methods to determine the effect of new technologies like Generative Artificial Intelligence (GenAI) on organizational performance, productivity, and efficiency of making a decision. The traditional enterprise forecasting models are prone to such problems as noisy records, redundant attributes, and high-dimensional data, which decrease the accuracy of predictions and elevate the complexity of the calculations. Considering this, the study presents a powerful deep learning-powered prediction model of enterprises that employs a combination of multiple stages of learning to enhance predictability and scalability. The preprocessing and standardization of raw enterprise data through Min-Max normalization is done first to improve the consistency of data and stability of the training. Next a dimensionality reduction method based on Autoencoders is used to extract the most important features and remove redundancy and noise. The detailed characteristics are further maximized with a Deep Neural Network (DNN) to establish complicated nonlinear connections and give the best feature weight. Lastly, Graph Convolutional Network based Prediction (GCN-P) model takes advantage of inter-enterprise relationships and structural dependencies to make valid adoption impact predictions. Through experimental assessment, it is established that the proposed framework has a greater predictive accuracy, reduced latency, increased throughput, better reliability and increased computational efficiency relative to the current means. The findings verify that the offered system is a scalable, stable, and efficient solution to the implementation analysis of the enterprise level of decision support and GenAI adoption.