FEATURE: STAYING ON THE CURVE II

Unpacking technology imperatives for chartered accountants

How will the maturation of data science combined with advancements in artificial intelligence revolutionise the role of the CA(SA)?

Last month we ventured into the future by touching on disruptive technologies and how they are helping to shape (or in many cases reshape) our profession.

In this, the second instalment, I intend delving deeper into the application of advanced data analytics and artificial intelligence to the accounting and auditing profession. By examining real world examples, perhaps we can better prepare ourselves for the sea change that lies ahead.

Comprehensive research already provides no doubt that rapid progress in our ability to derive actionable insights from analysing big data combined with breakthroughs in artificial intelligence pose a threat to a large spectrum of jobs, including those where the application of judgement is paramount.

It is this threat to even white-collar jobs that many would dismiss as absurd.

Indulge me, if you would.

Earlier this year, Fukoku Mutual Life Insurance in Japan laid off 30 experienced insurance assessors and replaced them with IBM’s Watson Explorer, an AI system that can calculate payouts to policyholders. The company believes it can increase productivity by at least 30% and see a return on its investment in under two years, resulting in annualised costs savings in excess of R17 million. Certainly not pocket change in any business.

IBM’s Watson can apply judgement to rules and legislation, enabling it to analyse and interpret structured and unstructured data, text, images, audio and video. It will be able to read tens of thousands of medical certificates, and through interfacing directly with medical service provider systems, factor in length of hospital stays, medical histories and surgical procedures before calculating payouts.

Based on our dependence on rule-sets and task-oriented characteristics, it is predicted that there is a 94% likelihood accounting and auditing jobs will eventually become automated. In reaching this conclusion, it should not come as a surprise that entry-level, routine and replicable accounting and auditing tasks can be or already are being automated. The Big 4 have all invested heavily in creating tools to reduce the amount of time spent performing low-level and low value-add procedures where time is a scarce commodity.

The threat that big data analytics will replace many of the tasks traditionally performed by accountants may be particularly relevant in the auditing context. For example, instead of relying on traditional sampling techniques to perform tests of details, automated processes could examine entire populations for unusual patterns and anomalies. In place of auditors sending out manual confirmations, the application of blockchain technology, synonymous with the emergency of crypto currencies, could enable automatic confirmations.

A consensus is emerging among academics that once access to big data analytic techniques becomes ubiquitous in business, users of financial statements will expect audited financial statements on demand, necessitating a shift from traditional sample based auditing to continuous ‘auditing by exception’, where data analytic techniques direct auditor attention on a real-time basis to instances where the data does not match the auditor’s expectations based on his or her knowledge of the client’s business.

Furthermore, in the future, public accounting firms may face competition for the provision of audit services from non-audit firms. Advances in data analytic and data visualisation techniques will make it easier for non-accountants with data-analysis backgrounds to obtain audit evidence and complete financial statement audits by applying data analytic techniques to big data. Within the auditing industry and academia, there is growing sentiment that if accounting firms do not exploit opportunities or neutralise threats made possible due to emerging technologies, existing firms such as Google or FinTech start-ups may seize the opportunity to enter the audit market. This would make the competition for the provision of audit services exponentially more fierce.

It is not hard to conclude that as technology advances further, more complex tasks such as business analysis, external reporting and auditing may become highly automated as a result of the routine nature of these tasks and the lack of machine inimitable skill requirements associated with these tasks.

So where does that leave our profession? In an era dominated by machine learning and big data analytics, are accountants and auditors to become obsolete? Or will there be a dawn of new opportunities for us to enhance our skills and create what I shall term ‘precision value’ (or pV for short) within our chosen accounting fields, disciplines and specialisations.

There is widespread discussion regarding the impact of machines on employment. Mental tasks have traditionally been seen as a competitive advantage to humans. However, the computer revolution continues to blur the line between physical and mental tasks. One needs look no further back than the late 1970s and the introduction of the modern spreadsheet to see how both the roles of book-keepers and accountants were fundamentally changed. What typically took many hours of manual compilation was replaced by an ability to generate calculations quickly and cheaply. Today, advancements in software and hardware have given rise to an expectation that machines can assist with judgement, not simply replacing repetitive tasks.

The need for judgement arises where decision-makers cannot describe, given a set of underlying facts and predictions, a perfect outcome as a result of remaining unknowns or uncertainties, yet must reach an opinion based on logical thought and comparison by utilising the data (or evidence) at their disposal. Let’s apply this definition in the context of an audit.

An auditor is required to reach a conclusion on a set of financial statements by obtaining sufficient, appropriate audit evidence. However, auditing is not an exact science. Put another way, the auditor will always apply logical thought to the evidence – or data – that has been obtained in order to predict a correct course of action. In this case, the correct course of action is synonymous with the most appropriate audit opinion to be provided. In essence, the auditor is giving a prediction as to the most likely state of his or her client’s financial records for the period being audited. Rules and regulations are followed, past experiences will be drawn upon, and ultimately the auditor recognises a pattern or trend that steers them to give one or another opinion. Again, there will always be an element of the ‘unknown’, but auditors are trained and monitored to provide their judgement so as to achieve what we commonly refer to as ‘reasonable assurance’. This is a measure of reliability upon which users of financial statements can base their trust and decisions.

Artificial intelligence techniques (especially deep-learning neural networks) combined with big data analysis aim to provide better predictions. The medical profession is currently employing these advances to improve diagnostic outcomes and hence patient treatment. By being able to better predict an underlying condition and eliminate false ones, doctors and specialists are continuously improving their judgement. From oncology to radiology, relying on advanced AIs such as IBM’s much-touted Watson, health care is undergoing a revolution. This is a prime example of pV in action for patients.

In the same manner, auditors and accountants can leverage these disruptive technologies to eliminate even more uncertainty, pin-point downside risks, and identify opportunities on their audits or in their professional roles in commerce, industry and the public sector. As I have reasoned above, judgement is a measure of prediction given inherent uncertainties and unknowns. As I alluded to in my previous article, it is only a matter of time before shareholders and stakeholders expect higher levels of confidence and assurance from the auditing profession.

I turn to the example of a hypothetical insurance company I recently shared with my students. One of the more significant risks arising in the insurance industry is the payout of claims to holders of policies which have already been terminated. That’s akin to throwing money away and reimbursing an uninsured event. It shouldn’t happen, but it does. Auditors could typically test for this occurrence by using a combination of substantive tests of detail and tests of control. Current auditing regulations would allow auditors to test a small, statistically valid sample and reach a conclusion. The alternative would see the use of advanced system and data CAATs (computer-assisted auditing techniques) to test the entire population of claim payouts against policy terminations and report on any and all misstatements in the financial records and weaknesses in internal controls. As an investor or stakeholder, which approach instils more trust and eliminates more risk? It simply cannot be the limited sample when the technology, hardware and expertise exist to eliminate more uncertainty, provide a better prediction, and hence allow for a more accurate formulation of judgement. This is a simple example but one that readily explains the dichotomy between legacy methods and future advancements.

The dilemma facing firms, and academia in particular, is how to teach and instil an acute appreciation for technology in each of the four key disciplines our students and trainee accountants are expected to master. To be blunt: we are behind the curve.

Our current focus on technical mastery of rules and routines that can be replicated and in some instances bettered by technology puts us at great risk. Only by ensuring students are exposed to foundational data analytics and AI techniques and concepts as part of their under and postgraduate studies will the future generation of accountants and auditors be able to remain relevant and safeguard our profession from redundancy.

By adopting this approach we will be able to reach a synergy between academia, the public practice firms and broader commerce. Academics cannot rely solely on the latter to instil, nurture and develop such skills after they have left our institutions of higher learning. If you think of the entrance of trainee accountants into their respective learnerships as a vertically integrated supply chain whereby SAICA-accredited institutions are responsible for providing beneficiated inputs to be included in a finished product, we have some work to do.

Some of our contemporaries in North America, Britain and Oceania have already embarked on a revision of core curricula for their students to prepare them for a world dominated by the pervasive use of technology and analytics in their accounting subjects. Comprehensive surveys and research has already been done to support this initiative. Given the South African context where not all of our learners entering university have had pervasive exposure to information technology and computers from a young age compared to the West, poses an even greater hurdle for use to overcome. I fully expect that SAICA’s CA2025 project will play a pivotal role in how we achieve this and I am proud to be a part of this initiative so that we remain relevant when compared to our global peers.

It is maintained that many professions consist of a bundle of tasks that collectively are not easily automatable and that accounting is one of those professions. Their belief is that data analytics and artificial intelligence will instead change task structures within the accounting profession and will provide opportunities for accountants to leverage their existing skills in conjunction with newly acquired ones.

Let me be clear: the underlying foundation of what truly makes a chartered accountant comprises the ability to undertake a lifetime of learning. Someone who is at ease with a continuous cycle of learning, unlearning and relearning will be able to reinvent themselves for the future.

This is not a new paradigm of thinking. No, dear reader, these tried and tested words of wisdom are ascribed to the pre-Socratic Greek philosopher Heraclitus, who said: ‘The only thing that is constant is change.’

At a university level, incoming accounting students are taught ‘accounting is the language of business’ and just as mastery of language allows one to understand and interact with native speakers of a language, mastery of accounting and its core disciplines grants the ability to interpret and understand concepts native to the business environment. This affords us the ability to think holistically about the information we are confronted with rather than responding in a predictable, algorithmic way.

And while advancements in artificial intelligence are allowing machines to provide vastly better predictions and, inter alia, improve applied judgement while learning from their mistakes, judgement does not automatically give rise to thought.

And thought, as we know it, has not been reduced to an algorithm or programmable code. Not yet.

It is this unique ability to think about what lies ahead that will keep us resilient in the face of adversity. True grit dictates we owe it to ourselves.

Author: Sven Wüsthoff CA(SA) is Head of Department: Accounting Sciences at Nelson Mandela University

References are available on request

Please rate this article

Relevance

Quality

Leave a Reply

Your email address will not be published. Required fields are marked *