Recently, the Economist published “What is an audit for?” and identified the current expectation gaps between what the consumers of a financial statement audit (investors, board members, lawmakers, courts, etc…) expect and what is being delivered. Examining these gaps provides tremendous insights into the imminent change in store for the Accounting Services profession.
I’ve taken the liberty to pull out a few excerpts and provide commentary.
Investors are also waking up to audits. They almost never vote against management’s choice of auditor. But last month over a third of shareholders at General Electric, an industrial conglomerate, voted against the reappointment of KPMG. Investors in Steinhoff are suing the company and Deloitte for $5bn for their losses.
CM: This is the critical broad trend to understand, the ultimate customer of the audit is the investor, BoD, and others seeking evidence of financial performance and more importantly, trends toward an outcome whether that outcome is positive or negative. Although Auditors will likely never speculate on an outcome, the data that they gather is a gold mine on which to apply A.I. and predictive analytics. The value of A.I. is driven by the quality of training data.
Robin Litjens from Tilburg University says there are several good reasons why failures may not always be detected. For one, a company’s books are so vast that audits can only realistically assess a sample of transactions in selected markets. Auditors hope that better data-analysis techniques should allow for larger samples and better anomaly detection. But for now, for large firms, looking at less than 5% of transactions is not unusual.
CM: The underlying assumption is fundamentally flawed. It assumes humans are required to “assess” or verify the sampled transactions with evidence (independently sourced and verified by a 3rd party). If you have access to quality data that can be used as evidence, computers are infinitely faster and more accurate. The real challenge is simplifying the process to collect the evidence and apply it to existing accounting records. This is the entire concept behind the new category of systems we see emerging called VFI (Verified Financial Intelligence) – a very specific type/application of A.I.
Similarly, auditors look only for errors that are “material” compared with profits or assets. The threshold is often in the range of 0.5% to 10%. These limitations might help explain why, according to the Association of Certified Fraud Examiners, auditors picked up only 4% of occupational fraud in 2017. Although some firms offer more forensic audits, they cost so much in time and money that companies choose them only if they already suspect wrongdoing.
CM: Automating the evidence collection process eliminates the need for sampling and isolates timing differences and non-cash transactions, or where the majority of the risk lives. The concept of “tolerable error” caused by sampling goes away, accuracy is improved, the risk is reduced. If you can automate the highest level of scrutiny applied, why wouldn’t you apply that up front and proactively in all cases?
Thank you for your interest. If you are interested in learning more about how VFI (Verified Financial Intelligence) can help with the audit process, I’d enjoy having a conversation.
CLICK HERE – to set up a time for a conversation!