As companies grow larger through mergers and acquisitions, the number of identities within their organizational structures increases exponentially. These identities stem from diverse sources and serve various purposes, creating a significant challenge in effective classification. Generally, identities can be categorized into four main groups: Employees, who comprise the internal workforce; External Users, including clients and partners; Service Accounts, used for automated processes; and Generic Accounts, which allow shared access to resources. To address this classification challenge, we propose a method that begins with identifying keywords associated with each identity category. This initial step is crucial for organizing identities. We then apply preprocessing techniques using lemmatization algorithms to standardize the terms, ensuring uniformity across different usages. Finally, we employ machine learning models to automate the classification process, improving the efficiency and effectiveness of identity management in large organizations. This comprehensive approach not only tackles the complexities of managing multiple identities but also enhances overall organizational security.