Artificial Intelligence. The “Watson” to the forensic auditors’ “Sherlock”?
Artificial Intelligence. The Watson to the forensic auditors Sherlock
Auditors have always been considered watchdogs, not bloodhounds.
The primary objective of auditing is not to detect frauds or fraudulent activities. Instead, auditors are meant to frame an opinion on whether the information in financial statements is materially presented in accordance with GAAPs. While one of their responsibilities includes intimating the management about presence of a material mismatch or fraud, the primary responsibility of fraud prevention and detection lies with the management.
Here comes forensic auditing (also called forensic accounting sometimes) into the scene.
Forensic auditors go beyond the financial statements and collect data from various internal and external sources to identify anomalies and collect audit evidence. They look through financial and statistical data and conduct interrogations and interviews with the company officials and staff to detect fraud and other illegal business activities. Every time an accounting scandal or fraud has been uncovered, the need and importance of forensic auditors have strengthened.
However, even forensic auditors face several challenges today. They have to deal with —
- The growing ‘expertise’ of fraudsters who are finding new, innovative ways of doing fraudulent activities, and
- Growing data.
With advancement in technology, the importance of AI software and products has become indispensable for forensic auditing and investigation firms to uncover frauds and deal with advanced issues.
Our today’s article discusses how Artificial Intelligence (AI) and its subset Machine Learning (ML) can empower the forensic auditors:
Big data technology
Big data can generate connections between two financial and nonfinancial information to discover potential patterns. It includes extraction of data from unstructured types such as texts, social media content, conference reports and interviews
Neural networks can model complex and nonlinear relationships in data to detect accounting fraud. They recognize trends and complex relationships which can hardly be noticed by other computational methods.
Match between two unrelated parties
AI compares data across various databases and systems — especially those where the data is never compared. For example, it will examine the suppliers’ payment transactions to check those instances where a supplier’s name, address or bank account matches with the details of an employee. Comparison among altogether different or unrelated datasets in a limited time frame is nearly impossible for humans. AI gives them added advantage to look at the matters in depth.
Variety of data systems and tools
Financial auditors have to perform audit checks on the books of accounts and other financial records maintained by the management.
However, for forensic auditors, the most arduous task is to deal with several data systems such as back office systems, customer relationship management systems (CRMs), social media, emails and mobile messaging applications for collecting data and finding patterns.
Besides analyzing the structured data, which can be located from the accounting system or an Enterprise Resource Planning (ERP), the forensic auditors also have to locate the unstructured data available on emails, WhatsApp and other channels of communication.
AI can locate the information, identify the patterns and sort the data into manageable groups for human analysis – in less than one-tenth of the time that a team would take.
Materiality vs suspicion
An external or internal auditor gauges the effect of each transaction based on the materiality approach. They tend to avoid checking an area or a set of transactions which don’t have a materialistic impact on the company’s financial statement.
However, a forensic auditor always looks for even a small transaction that looks suspicious because it could be a gateway to a huge accounting fraud.
AI goes beyond the value of a transaction and applies tests to every transaction or event that comes under its radar.
Intuition vs data-driven approach
Human auditors have always relied on intuition-driven approach in finding the samples and conducting audit checks, based on their expertise and industry knowledge.
AI is more focused on progressive continuous learning which can detect the new, previously unseen fraud tactics quicker than their human companions. Because of its augmented cognitive capacity, it learns, stores the information, and becomes smarter after every transaction or process.
As against the popular misconception “AI will take over the humans”, this article strengthens the fact – forensic auditors will get empowered, not replaced. AI will give more power and time to the auditors to focus on areas which require their consideration and focus.
With benefits like ability to extract data from various structured and unstructured formats, identification of unusual patterns, focus on data-driven approach, and improvement in audit quality, forensic auditing will thrive.
What’re your thoughts?