US Auditors Risk Losing to AI by Ignoring Sustainability Data

We stand at an inflection point where regulators, the C-suite, and especially the accounting profession can seize an opportunity to drive a relevant and more complete view of each enterprise in the marketplace.

Financial reporting with double-entry bookkeeping and barely integrated sustainability metrics make the US and other jurisdictions vulnerable to incomplete information.

In just one of many examples where organizations fall short, most current financial reporting lacks balance sheet information about what the audit profession calls constructive obligations, such as promises related to a company’s sustainability claims. While International Financial Reporting Standards require the entity to make a financial statement provision for such programs it plans to roll out—for example, a promise of net-zero emissions—most don’t.

An audit profession broadly steeped in US methodologies stands ill-equipped to opine on the European Union’s Corporate Sustainability Reporting Directive, which requires integrated reporting of financial and sustainability information from about 10,000 companies, including those headquartered outside the EU, with many in the US.

In addition, the audit profession faces strong competition from rapidly emerging data analytics and artificial intelligence, as well as from the non-audit firms that can better leverage those technologies.

In early March, Japan issued sustainability disclosure standards, initially affecting the Tokyo Stock Exchange’s prime market-listed companies.

As the EU, Japan, and other nations broaden their publicly reported financial and corporate health disclosure requirements to include an organization’s waste output and how they dealt with it, how can anyone know if that data will be complete? We can’t.

Many organizations report financial outcomes—and markets make investment decisions—based on analytics, systems testing, and related measurements. This helps independent auditors form an opinion based on sampling across an enterprise.

Most use US generally accepted accounting principles or a derivation of them, because those organizations are based in or do business with the US market.

The auditor opines on the correctness of the assets and the completeness of the liabilities side of the ledger, known in the markets as auditing for overstatement of shareholders’ equity. The objective is to protect new investors from putting their money in a company that tries to look better than it is.

Smart automation, particularly artificial intelligence and big data analytics, “will sooner rather than later (due to lagging audit innovation) take over all human audit work as based on complete data sets,” the Canada- and Netherlands-based publication ComputationalAuditing.com recently reported.

Organizations that focus on machine-learning and analytics have become positioned to achieve the more complete view of an organization, as opposed to the auditor focused on sampling. Yet the auditor stands uniquely qualified and trained to understand how the non-financial corporate information relates to the financial statement. Therefore, they remain best positioned to deliver an “integrated” holistic audit with insight across the enterprise.

Dutch audits assume an understatement of net revenue, intentional or not, by diving deeper into the functions of producing a product or service. This allows shareholders to remain confident in their dividend payouts and governments won’t miss out on tax revenue.

While the US audit model “follows the money”—analyzing cash flow for overstated revenue, the Dutch audit “follows the stuff”—detecting unstated or withheld revenue by focusing broadly on product development and the processes that precede cash flow.

So it’s no wonder the US methodology based on samples became the preferred audit, while the Dutch completeness goal hurt broader adoption of the methodology because it was more expensive to go beyond sampling. Now, machines can perform that heavy lifting.

The Dutch system’s analysis can add disclosures on the viability and sourcing practices of supply chains, an organization’s energy usage, workforce availability and skills sets, corporate governance and even the strength of the board, as well as other measures rolled up under an organization’s environmental, social, and governance capabilities. This would include the constructive obligations and other items.

Once you have a relevant complete picture, each area likely can be better managed, and possibly even strengthened.

This could become a major step toward more fully understanding a company’s revenue, its supply chain input and output, and a cradle-to-grave investment allocation for each consumer product, as well as having the associated costs and processes to offer a better understanding of the carbon and other pollution footprint.

Without this marriage of auditing and assurance, the current US approach raises the risk of undetected completeness failures, potentially imperiling financial markets and likely raising audit firms’ liability risks. Integration offers a superior solution, based on the very powerful double-entry bookkeeping principles that non-auditor firms can’t offer.

Otherwise, the non-auditor professional services firms that advance AI and big data analytics may become the go-to for financial analysis to provide product-based completeness assurance and enterprise-wide insights across the supply chain.

There seems to be little stopping those non-audit consulting firms from expanding their capabilities to throw in the financial opinion as an add-on service—for free.

This article does not necessarily reflect the opinion of Bloomberg Industry Group, Inc., the publisher of Bloomberg Law and Bloomberg Tax, or its owners.

Author Information

Eric Israel, a chartered accountant and certified fraud examiner, co-founded KPMG’s Global Sustainability Practice, helped co-found the conflict minerals practice at PwC, and was US director for the Global Reporting Initiative.

Write for Us: Author Guidelines