AI Based Transformation Projects-ML and NLP Managed Audit, Control, and Governance (ML_ACG) - the Basics

This article focuses on Artificial Intelligence (AI) based Transformation Projects, in which Machine Learning (ML) and Natural Language (NL) Processing (NLP) are used to interface, integrate, and manage enterprise’s Audit, Control, and Governance (ML_ACG) Framework (ML_ACGF). It also promotes an In-House Implementation (IHI) approach for such a holistic framework. Knowing that today the fields related to audit-analytics, and the use of AI and NLP, have become a strategic factor for transformation projects. Such projects even revolutionize the manner in which auditors extract and use insights that are sourced from huge data-volumes and then used by specialized frameworks, like the ML_ACGF. The ML_ACGF, uses various AI Subdomains or fields (AIS), in order to: 1) Enable accurate analysis, interpretation, and generation of human-language based reports; and also to improve risks’ assessment processes; 2) Discover hidden insights in documents, data-sets, specialized audit-medias, design-documents, source-code, emails, and other contents’ formats; 3) Automate audit procedures by interfacing NLP, Machine Learning (ML), and Natural Language Generation (NLG) modules, where ML learns from historic data-sets to find hidden patterns and to enable the prediction of possible important projects’ risks, knowing that predicting such risks is a very critical subject, because transformation projects have an eXtremely High Failure Rates (XHFR), which is about 95%; 4) Project’s architecture and related models, are used to instantiate solutions by applying standard Enterprise Architecture (EA) and audit methodologies; 5) Use decision-making processes to mitigate critical risks; and 6) Present interfaces’ and integrations’ technics for existing market’ auditing frameworks like the Control Objectives for Information and Related Technologies (COBIT), Committee of Sponsoring Organization’s (COSO) Enterprise Risk Management (ERM), and other. For this article the author: 1) Uses his research framework that is based on a mixed-method and a Polymathical approach, which is specialized in supporting transformation projects; 2) Implements and presents a Conceptual Proof of Concepts (CPoC), which illustrates a concrete case and solution; and 3) Presents a list of conclusions and recommendations on ML_ACG’s feasibility and advantages, where the main goal is to provide models and concepts that support and implement ML_ACGF’s different features, and not just to use quantitative analysis to show commercial opportunities.