Classifying 1.5 Million Columns: An Automated Multi-Engine Approach to Enterprise PII Governance
Year: 2026
Type: conference-paper
Abstract: The rapid growth of enterprise data repositories has significantly increased the challenge of identifying and governing Personally Identifiable Information (PII) across large-scale database environments. Organisations often manage millions of database columns distributed across heterogeneous systems, making manual classification impractical, time-consuming, and error-prone. This paper introduces an automated multi-engine approach for classification of 1.5 million database columns that can help enterprise PII governance and regulatory compliance. The suggested method combines several detection engines such as keyword-based detection, pattern recognition, semantic similarity analysis, metadata analysis and machine learning classification to accurately detect sensitive information. The individual engines evaluate the column properties, and a confidence score is computed. Then the scores are fused together using an ensemble decision mechanism to form the final classification. This framework allows for easily and efficiently identifying PII elements that can be scaled to multiple devices and users, including names, emails, phone numbers, identification numbers, financial data, and customer records. Experimental results show high classification accuracy, higher precision and fewer false positive cases and better scalability in handling large enterprise data sets. The recommended solution enhances data privacy governance, helps meet compliance mandates, and offers enterprises a smart way to discover and protect PII in an enterprise-wide manner automatically. Keywords—Personally Identifiable Information (PII) Classification, Enterprise Data Governance, Automated Data Discovery, Multi-Engine Classification Framework, Privacy Compliance and Data Protection.
Authors: Mayank Sethi, Pranay Kammari, Hanish Chalicham
Institutions: Wichita State University, University of Connecticut
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7203480134
