Data mining, a process of analyzing massive volumes of data, is profoundly transforming auditing. Based on the comprehensive analysis of accounting and operational data, this approach makes it possible to gain efficiency, reinforces the reliability of controls and provides a better understanding of how businesses operate. Two partners at Nexia S&A, Bettina Koegler and Charlotte Jansen, explain how this evolution is creating shared value between auditors and customers.

What are the benefits of data mining?

Bettina Koegler : This is a real gain in efficiency and comfort, for us as well as for our customers. We now process all the data rather than a sample, which makes controls more reliable and reduces repetitive tasks. The time saved is reinvested in the analysis and the exchange with the customer. Data mining paves the way for a benchmark between companies and sectors on their operating methods that allows us to compare practices and share areas for improvement with our customers.

Charlotte Jansen : Data mining offers us a vision that is both very detailed and very global of the functioning of a company. By modeling your flows — turnover, receivables, payments —, you better understand your business, its risks and its performance levers. The analysis is no longer based on random tests: we have an exhaustive vision that makes it possible to target risk areas, to adapt the audit with greater added value and relevance, and to focus our analyses where the real challenges lie. It also allows our customers to have a more macroeconomic vision of certain flows. Seeing an overview makes it possible to better identify unusual flows and secure them, from an operational point of view and from a legal/accounting point of view.

Where are the limits of this approach?

Charlotte Jansen : The main limitations come from the data itself: its quality, its coherence, its dispersion between several systems. The Accounting Records File (FEC) provided a standardized basis, but not all the information is yet uniform across companies or tools. Identifying the right data, making it usable and understanding its flows requires real work beforehand. The challenge is also not to get lost in the mass of information: you have to know how to target what is relevant and useful for the audit.

Bettina Koegler : Yes, it all depends on the degree of mastery and structuring of data within companies. Information often has to be cross-checked and framed in order to be able to use it. This requires perspective, but also real expertise. Data mining therefore requires more experienced auditors, capable of interpreting analyses with a critical eye and of relating the results to the reality on the ground. The added value lies above all in understanding the data in order to make a relevant analysis.

How to make data mining an effective audit tool?

Charlotte Jansen : For data mining to be effective, you need, above all, reliable and well-structured data, as well as a customer who knows its processes and flows. The analysis is always built with him, in discussions with the chief accountant, the CFO or the financial teams. This joint work makes it possible to understand the exceptions, to make the results more reliable and to better link the data to operational reality.

And how is this approach put into practice at Nexia S&A?

Bettina Koegler : Data mining allows more efficiency thanks to an increasingly broad benchmark for more and more detailed analyses. We have already developed our own analysis tools for cycles, with more flows such as turnover or stocks. The arrival of electronic invoicing will further enrich the data: it is not an immediate breakthrough, but a continuous evolution that prepares for real transformation, driven by intelligent agents.

‍

ELECTRONIC INVOICING: A MAJOR ADVANCE

Electronic invoicing is a real turning point for auditing: it will standardize a large part of financial data, beyond what the FEC already allowed. The information will be more detailed and consistent regardless of the channel used, making it easier to use it. For auditors, this is an additional advance: the data will be more reliable and immediately usable, even if the volume to be processed will require strengthened tools and analysis capabilities.

BETTER DETECT ANOMALIES

The vision of both overview and detail makes it possible to quickly identify transactions that are out of the usual pattern. An “anomaly” is not necessarily an error, but an exception to be understood. For example, the analysis can reveal receipts going through several accounts before being documented, or the absence of contracts on certain flows. The questions identified allow our customers to simplify processes and strengthen controls in risky areas. The tool refines the understanding of accounts and makes it possible to target the most exposed areas.

DATA MINING

Data mining can be applied to accounting data and operational data. It is enough to define a usual pattern of operation and then to identify the exceptions. Data analysis and the use of AI to exploit them are the challenges of the evolution of our business. In the future, the auditor will be assisted by “Big Data” analysis tools, AI and process mining (automatic control point analysis). The challenge will be to identify the right data and the right flows to target the audit effectively.

‍

Definition of data mining: Data mining consists in automatically exploiting large volumes of data to extract trends, identify operational flows and model the functioning of a business.