ÂÜÀòÉç¹ÙÍø

We’ve heard more and more over the past several years about how artificial intelligence (AI) could be used in internal audit.

A of 800 executives found that 75% of those executives believe 30% of every corporate audit will be performed by AI by 2025.

A lot of organizations are investing in AI because it allows them to entrust more tasks to machines, freeing up their personnel to focus on strategic priorities. Efficiency is the name of the game when it comes to AI.

But how could AI impact internal audit, specifically?

In 2017, the Institute of Internal Auditors (IIA) released a paper titled , which helped lay the foundation for how to perform audits on an organization’s AI functions. So there is guidance for auditors who were questioning how to deal with the rise of AI technology within the organizations being audited.

But we’re going to focus on how internal auditors themselves could use AI, and whether AI could replace the data analytics that sophisticated internal auditing departments and outsourced providers currently use.

How could AI affect the use of data analytics in internal audit?

Larger companies with strong, established data analytics programs in their internal audit department are starting to look into or are already beginning to utilize AI. Companies with much smaller, leaner internal audit departments and without the budget to invest in data analytics and AI — and that may not have the amount of data needed to leverage AI — aren’t finding it worthwhile to look into using it.

Because of the data, budget and time requirements, the organizations that could most benefit from using AI in internal audit are the ones who already have a strong data analytics component to their audit function. Without data analytics as a foundation, you cannot build AI into internal audit.

Currently, these more sophisticated organizations are using data analytics to assist in developing their audit plan, enhance their enterprise risk-assessment efforts, and better scope their audits. Then they utilize data analytics to quantify and determine the severity of observations identified during the audit.

So what does bridging the gap from data analytics to AI look like? It would start with correlating the data to develop risk- and score-based assurance models and automating transactional testing. AI integrated into data analytics has the potential to ingest and analyze 100% of transactions instead of just a sample, and it could identify and flag anything that doesn’t seem right so that auditors can look deeper into those flags.

Using AI in transactional testing means auditors could spend far less time performing and reviewing transactional testing and instead refocus their efforts on qualitative questions and inquiries. Essentially, AI can build off of data analytics’ current capabilities, increasing the amount of heavy lifting that machines do in internal audit and allowing auditors to focus on more impactful activities within the audit. Because 100% of transactions are analyzed, AI also decreases the risk related to giving a final opinion on the nature of the financial statements. Plus, it decreases expenses related to the audit.

How do we get there?

Before any internal audit department begins using AI integrated with data analytics, there are barriers to overcome.

First is a big risk to consider: human bias. AI is built by human beings, and the risk of those human beings inadvertently imbedding bias into the technology is higher than most people are comfortable with.

The predicts that 85% of AI projects through 2022 will deliver erroneous outcomes because of bias in data, algorithms or development teams.

When it comes to internal audits, accuracy is critical, making bias a pretty big obstacle to overcome.

Yet an even bigger barrier to overcome is data quality. Being able to even use data analytics requires a large amount of time cleansing data and making sure you have complete and accurate data. Similarly, AI needs vast amounts of accurate, complete data to function correctly. This is the main reason why larger companies with bigger budgets and huge amounts of data are able to use data analytics and begin looking into AI’s capabilities. Gaining accurate data is one of the biggest hurdles to overcome.

Another significant barrier that many internal audit departments encounter is a lack of the right talent and skillsets required to effectively build out a data analytics function in their department. The ideal individual has a combination of technical and digital, IT, accounting, audit, business, people and leadership skills, which is a hard combination to find. Oftentimes, they will co-source or outsource that function of their internal audit department if they have the budget to do so.

Finally, leadership itself can be a barrier. Many organizations understand the benefits of using their own data to gain insights and make proactive decisions, but in reality, it takes strong leadership from the internal audit department on up to the C-suite to make data analytics a reality for a cost center like accounting. Leaders who understand data analytics’ benefits and requirements and are proponents of it can help greenlight and build out an effective data analytics program for internal audit, as well as the rest of the business, and potentially integrate AI into their data analytics program at some point in the future.

Is your internal audit ready for AI?

Some people are asking whether AI can replace data analytics in internal audit and perhaps even auditors themselves. The short answer is no. It cannot replace the experience and judgment of auditors. But what AI can do is build on top of a solid data analytics foundation and begin taking on even more of the heavy lifting for internal auditors by automating and accelerating large and complex data tasks and analyzing larger volumes of audit-relevant data in order to derive better insights and audit findings.

Whether AI will actually be doing 30% of every corporate internal audit by 2025 remains to be seen, but given how fast the technology is progressing, we wouldn’t entirely rule it out. AI is now, not the future, and the sooner that internal audit departments embrace and implement data analytics and AI functions, the better off they’ll be.

Related reading:

Are you following the IIA standards for your internal audit practice?

What is continuous auditing and how can you leverage it?