Recent progress in digitization has turned business-as-usual auditors into visionaries. Auditing today harnesses robotics, automation and cognitive technology (or artificial intelligence (AI)), developments so new that even their names would’ve moved executives to scratch their heads just two decades ago.
And as recently as the early-2000s, many, if not most, professionals would’ve been circumspect if asked to envision where we’ve now arrived. Sea changes in auditing usually take generations. Not so today.
“It’s an incredible time to look back, staring down from today, and then look forward,” says Bill Tomazin, Regional Managing Partner, West and Managing Partner of National Audit Solutions at KPMG U.S. “Over the past three years, we’ve moved quickly into digitizing the audit and helping our auditors obtain structured data from our clients—and using technology to help them make better sense of that data. It’s really exciting because it helps our teams make better decisions and raises the quality of the audit.”
And as the traditional processes of the audit profession and financial reporting turn over, it’s critical for firms to embrace advanced technologies, now. “The development and maturity of cognitive technologies and the ability to mimic human judgment over the next three to five years will be a game changer in terms of the audit,” Tomazin predicts. These technologies will advance like never before. “That’s where real audit transformation begins to take place—when we gain insights and are able to make better decisions through data and capabilities we didn’t have before,” says Tomazin.
Look around in 2018 and you’ll find many auditors following Tomazin’s lead as they view the latest developments in AI with interest and enthusiasm. But the profession is also grappling with some uncertainty. A moment of change has arrived that promises unprecedented breakthroughs at breakneck speed—or as many in the industry call it, “digital speed.”
In terms of audit quality, that means figuring out first and foremost where the cognitive tools and technologies belong—and it’s not at the front of the priority list. For, as with any computer technology, the output depends on the robustness of the input: in this case, data from business clients.
“We have adopted a data-first philosophy,” Tomazin notes. “We focus our efforts first on understanding the data that an organization has available that we, as their auditor, can leverage in understanding risks and gathering audit evidence. Once we have a vision of available data, we then focus on the most effective way to procure and position that data for various use cases on the audit. We then apply the appropriate tools and technologies to automate, analyze and visualize.”
Meanwhile, the profession remains in the early innings in terms of figuring out how best to leverage AI. What’s more, artificial intelligence is an umbrella term that covers many technologies developing at different rates: a broad category that reaches from natural language processing, which extracts attributes from data, all the way up to deep learning and neural networks that use sophisticated techniques to make predictions.
Extracting Data For Exceptional Results
How, then, to break down those subcategories into developments to watch for in the months ahead? First, look for pronounced movement on the data end of things. As Tomazin puts it, artificial intelligence is now coming into play to bring together data, analytics and the human decision-making process. Natural language processing techniques and the ability to read and process unstructured data are also gaining traction, working as companions to the human auditor.
This is welcome news for auditors because AI can comb through reams of data faster and more precisely than any person or team of people could. For example: AI can ingest large numbers of leases, read contracts or review journal entries and extract data for use by the auditor.
Tomazin describes one example: “We’ve been collaborating with IBM Watson to create artificial intelligence that works with commercial mortgage loan files. This involves the ingestion of a substantial amount of unstructured data, the identification of specific document types within the credit files and extraction of key attributes from those documents, along with detecting cognitive patterns that learn and mimic human judgment. These capabilities will augment our auditors’ judgments and decision-making process and ultimately provide better, more effective audit evidence.”
It then comes down to repeating this procedure so that AI gets better at the task, a concept known as machine learning. AI takes data, applies algorithms to it and then makes predictions on what the outcome will be. The more data AI sees with successive cycles, the more correlations it makes and the better its predictive capabilities get.
AI Processing Meets Auditor Professionalism
Experts see much of the forward motion on AI and data making headway in the next 12 to 18 months. Meanwhile, leaders across business fields are speculating about what happens to humans once AI hits critical mass, both in ubiquity and processing power. Some headlines have been dire. With auditing though, the profession demands many prized skills you simply can’t teach a machine.
Foremost among these are careful planning and sound judgment to reach conclusions based on the insights machines produce. The foundation of a high-quality audit ultimately begins “with audit objectives, the data available and determining what would help us best in designing the strongest audit procedure,” says Brian Foster, Audit Partner at KPMG U.S. So, while AI may be the auditor’s ally in enabling specific audit procedures, “we don’t look at the tech first,” continues Foster. That is, auditors play the key role in designing and executing a high-quality audit, with the technology following their lead; not the other way around.
Meanwhile, other areas of AI still need time to mature. With deep learning techniques and neural networks in particular, the profession is a bit further away from adoption because audit trails are needed to satisfy various constituencies, from clients to regulators and the auditing profession as a whole. Yet major innovations on that front could be years rather than decades away. As a firm, KPMG is investing in “supervised learning.” Think of this as machine learning that develops under the aegis of highly skilled auditors adept with data and analytics, steeped in leading practices and boasting sound professional judgment.
With these advances in technology, the next generation of auditors will need a new skill set. “They’ll need an enhanced portfolio of skills,” Tomazin notes. “Procuring and digesting data needs to be a core competency. So how do we recruit for and build that skill set?” Part of the answer lies with KPMG’s Master of Accounting with Data and Analytics program, which is now at nine universities.
Speaking of teaching, moving ahead also means using KPMG’s intellectual property to teach computers how to make certain predictions. Then auditors can determine whether AI-enabled programs reach similar conclusions to those based on historical data, and whether that data can assist in predicting outcomes.
To that end, KPMG and IBM Watson’s alliance reflects KPMG’s commitment to reinforcing its position as a market leader in the use of advanced technologies to support its services. The partnership between KPMG and IBM, which includes a focus on auditing services, builds on several successful KPMG initiatives designed to demonstrate the promise of cognitive technologies across a range of business and financial functions. “By applying Watson, KPMG is taking a forward-looking approach to extending its capabilities, helping professionals and clients gain new insights from critical enterprise information,” says John Kelly, Senior Vice President of Cognitive Solutions and IBM Research at IBM.
“The most exciting thing to me professionally is that for the vast majority of my career, an audit has been a pick-and-shovel manual effort,” says Tomazin. “There were no enabling technologies to help support high-quality audits other than sheer force. Now, it’s people, technologies and platforms. It makes the look and feel of an audit 180 degrees different from the audit of just five years ago. And to me, that’s really, really powerful.”
