Recent studies reveal that between 40% and 45% of activities currently carried out by labor force can indeed be automated; however, no definitive figures exist on this topic¹.  The main incentives for implementing this change are based on the fact that new technologies are being more and more able to imitate the human activity, taking over certain tasks with more speed and precision than persons themselves, with investment levels far from unreasonable.     

In this sense, Robotic Process Automation (RPA) has been gaining ground as an efficient way to automate repetitive and labor tasks in a variety of business areas, such as: finance, accounting, technology, legal, human resources, compliance, etc.     

The audit function is no stranger to this new trend since implementing RPA solutions can make such work more effective and efficient, allowing to augment the audit scope (e.g., statistical sampling vs. total population) besides automating more routine procedures (e.g., simple matching, confirmation circularization, arithmetic tests, etc.).  This, in turn, allows "freeing up" time to focus on more strategic and value-added activities such as writing a report that requires a deeper analysis, as well as professional judgment, which would not be appropriate to process through RPA solutions.

In any event, it is of paramount importance to identify which activities or tasks are susceptible of being replaced by RPA solutions, which typically will include those of a routine nature, labor intensive, prone to human errors and controlled by rules (that is, not subject to judgment), and for which the information that gives them origin is available and easy to read.

The application of RPA solutions successfully in the audit function involves a 3-step process, described below:    

  1. Understanding the processes – The audit processes that more rapidly would benefit from RPAs are those involving defined audit tasks that are repetitive, take a lot of time and do not involve the auditor´s judgment. The foregoing entails an ample knowledge of the business in particular and the industry, as well as the typical inherent risks of the latter.   Once the auditor has determined that a process is a good option for implementing RPA, the next step within the understanding of the process is to divide the audit tasks in small sub-steps that may be "interpreted" by the RPA solution (what may be intuitive for a human user, not necessarily will be for an automated application). 
  2. Standardization of the auditable information (Audit Data Standardization or ADS) – To be run as planned, RPA audit solutions require consistency in all data fields.  Data related to the audit may come from different sources (e.g., ERP systems of each client are different, process their transactions in a specific way and generate custom reports).  In this sense, data fields included in the reports requested by the auditor, which include the same information, may be different among themselves. Consequently, it is imperative to have ADS for each process that will be replaced by RPA solutions. A good ADS template or format will extract the relevant data necessary to execute the audit tests and, with this, guarantee that the RPA solution will accomplish its intended objectives. 
  3. Perform the audit tests based on RPA –  The successful programming of the audit tests as rule-based algorithms will allow the execution of such tests with a simple click (import, read data, perform audit tests and export results).                                    
    In any event, interpretation of results, which is the delicate part and from which a great part of the audit´s work is based on, continues to rest in the auditor´s judgment. RPA solutions support the auditor´s work for obtaining results in a faster fashion, with more objectivity and over an enlarged scope, in addition to providing elements to produce high-value observations to the business of our clients, but in no way will replace the ability (and responsibility) of the auditor for expressing his opinion.  

________________________

1 MIT´s article "Every study we could find on automation will do two jobs, in one chart", consulted on July 1, 2019, states that Oxford University determines that 47% of jobs in the USA could be lost as a result of automation, while McKinsey says that between 800 and 900 million jobs worldwide could be lost, at the same time that a study by the OECD suggests that 9% of jobs in the 21 member countries could be automated. Consultation made at: https://www.technologyreview.com/s/610005/every-study-we-could-find-on-what-automation-will-do-to-jobs-in-one-chart/

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.