Artificial Intelligence in Performance Management: Toward Automated and Productive Workflows

This paper introduces a performance management framework driven by AI that will help establish automated and productive workflow. This system proposed combines the Z-score normalization and fairness reweighing to remove bias and provide a balanced representation of data, followed by the mRMR feature selection to retrieve the most relevant and nonrelevant performance indicators. XGBoost classifier with the SHAP explainability can be used to make the correct and explainable prediction of the outcomes of employee performance. The framework applied with the PyCaret integrated suite of ML-XAI guarantees the efficiency of the processing, reproducibility, and interpretability of the framework in a single environment. The best experimental tests showed better outcomes in terms of fairness, precision, and consistency of decisions, as opposed to conventional ways of appraisal. The results indicate the promise of explainable AI in creating fair assessments, improving managerial understandings, and improving the efficiency of a company in general. The given approach creates a basis on the next-generation performance management systems that operate under the power of automation, transparency, and data-based decision-making.