What can AI do for Auditors?

Author: Ethan Joffe, Staff Software Engineer

Remember the good old days when auditing meant drowning in a sea of paper, using a calculator until your fingers went numb, and drinking enough coffee to fuel a small spaceship? Welcome to the 21st century, where AI is ready to rescue you from the drudgery of traditional auditing. In this post, we’ll explore how Artificial Intelligence can supercharge your work and why now is the time to let it become a part of the way you get things done.

AI has been around for decades, so why is now a good time for auditors to lean in? How can we leverage AI to transform the auditor’s role from number cruncher to trusted advisor, adding value to companies beyond traditional responsibilities?

The Promise of AI

Did you hear about the AI that caught a fraud scheme by noticing a pattern no human could detect? Or the one that analyzed thousands of transactions in the time it takes to microwave popcorn? These aren’t just fun anecdotes; they’re stories showing how AI is changing auditing.

With AI as a work partner, large volumes of data will be quickly analyzed. Key information will be more easily identified providing insights into potential risks and compliance issues. The accelerated pace will leave time to spare for next level contributions, opening up opportunities to engage in more intellectually stimulating and value-added activities.

“It’s elementary, my dear Watson”

Bill Gates recently observed “We now have a better sense of what types of jobs AI will be able to do by itself and which ones it will serve as a copilot for.” (1) AI’s great at picking out odd patterns, but it’s up to us humans to figure out the ‘whys’ behind them. It’s this human touch in the auditing process that keeps our role irreplaceable. With the efficiency gains from AI, we will now have time to dig into the nuances of these anomalies, using expert judgment and experience to uncover and understand the context behind them. This detective-like aspect of auditing is far more engaging and challenging than routine number crunching.

I have an auditor friend named Julia who works for a large tech company. She also happens to be a fan of AI, and has been trying to incorporate it into her work. While auditing corporate waste materials, she uploaded the company reports into an AI clustering algorithm, a technique used to group similar data points together to help discover patterns. Julia visualized the data, and noticed a weird cluster. She dug down deep into the source of the cluster, and realized there was a small but repeated classification error, buried in mundane reports. Turns out, the error was serious enough that it could have led to millions in potential fines. When I asked her how she felt about having to do all that work to find the error, she told me this was the part of her job she loved the most.

Julia has a passion for AI, which gave her the tools needed to detect and drill into the issue, but what about the rest of the audit community? Luckily recent advances in AI have opened up these superpowers to the rest of us.

Decoding Data, Crafting Strategic Narratives

As auditors, we are constantly setting up controls to provide actionable insights in available data. A deeper level of analysis allows us to interpret data and better understand the business implications and provide strategic advice to stakeholders. This forward-looking approach is crucial for helping our organizations navigate an increasingly volatile business environment by not just identifying current risks but also predicting future risks.

We have probably all heard about large financial players grappling with the complexity and volume of data involved in risk assessment. Traditional methods are time-consuming and too slow to be useful for identifying emerging risks. Starting in 2010, AMEX has employed machine learning techniques for a wide range of use cases, most notably in fraud detection. (2) In 2017, JPMG deployed their LOXM system, to execute trades for clients with maximum speed and optimal prices, based on lessons learnt from billions of past trades. (3) By leveraging AI to analyze the data, identify patterns, and predict potential risks, these large corporations are able to proactively mitigate risks before they materialize.

These institutions have been able to implement AI solutions because of their vast resources, but with recent advances in AI, this type of functionality has been commoditized enabling the rest of us to also create these systems. If you haven’t already, you can expect to notice many organizations leveraging AI in their products. SaaS providers, like AuditBoard, are starting to use AI to enhance products and to provide even more value to customers.

The Future is Now

Artificial Intelligence has been around for decades, with isolated utility available to only the largest organizations. But we live in a time where we have witnessed game changing advances in the availability and capabilities of these systems. In particular, we have seen the explosion of Generative AI, where computers generate original content, largely with the unveiling of chatGPT.

In 1950, Alan Turing proposed a test to assess a machine’s ability to exhibit intelligent behavior indistinguishable from that of a human. The original format, described by Turing, involved what he called the “Imitation Game.” This involved three participants: a computer, a human, and an interrogator. The interrogator, isolated from the other two, communicates with both and must decide which one is human and which one is the computer. If the interrogator cannot reliably tell the computer from the human, then the computer is said to have passed the test, demonstrating human-like intelligence.

“Today, product and engineering teams drive AI investment, but as solutions evolve to deliver more value, we expect generative AI investment to increase across departments.” (4)

Enterprise adoption had a marked uptick in 2023 and the trend is certain to continue.

Over the last few years, Generative AI has gone from being laughably obvious to indistinguishable from human content, while allowing it to be custom tailored to any audience. Advances in the speed and storage capabilities of computers and the scale afforded by cloud computing allow AI systems to generate content at a seemingly impossible rate. The nature of these algorithms and scale at which they run have made them available to everyone, guaranteeing their exponential adoption.

Quality of content

The “uncanny valley” was a term coined to describe the feeling of unease in reaction to content which is almost, but not quite, like what a real human might produce. Modern AI systems have transitioned beyond this stage and since the dramatic entrance of chatGPT onto the scene in the last year, has become indistinguishable, and in many cases preferable, to the musings of us mere mortals. Beautiful images, clever prose, and analysis at scale are just a few of the examples splayed across the pages of the media in recent months. And yet we should understand that these systems are not yet intelligent. It is time for the Turing Test to be reformulated, and companies like Google and OpenAI are doing just that. In the meantime, we can incorporate this high quality content into our daily work.

Speed is a Superpower

Two time Nobel prize winner Linus Pauling once said “The way to get good ideas is to get lots of ideas and throw the bad ones away”. Large language models like chatGPT are capable of helping us generate and refine ideas. It’s like having an all knowing oracle, where all we need to figure out is how to ask it the right question. Get lots of ideas, and throw out the bad ones.

Universal Access

At this point we have all played with or at least heard about how chatGPT can do amazing things. From AI rap battles to corporate summaries, it seems we are limited only by our imaginations. This commoditization of AI in the last year has created a seismic upheaval across many industries, in particular and early on for those focused on content such as media companies, writers, and artists.

The significant increase in content quality, the speed of delivery, and universal access will lead to a sea change in the way audit is about to get done. Now is the time to lean into these new capabilities and up our auditing game.

Fear Not the New

With all the promise of AI, there is still a large amount of resistance from enterprises over utilizing AI. There are valid concerns over data governance, accuracy and reliability, and the associated costs of AI. These concerns are actively being addressed, paving the way for smoother AI adoption.

One critical new worry is that corporate and personal data are being misused in training AI systems. This concern should be quelled with robust enterprise service guarantees along with broader education on the topic. In addition, regulatory bodies are quickly putting safeguards in place, such as the White House executive order on AI (5) and the European Union’s AI Act. (6) As far as the reliability of the results, human auditors must play a crucial role in monitoring and verifying AI-generated work. Since AI has now become commoditized, any costs are clearly outweighed by the benefits. When seamlessly integrated into audit processes by providers like Auditboard, AI is not just a cost but a smart investment.

Final Thoughts

Let’s talk about future-proofing your career. The world is moving towards AI integration in almost every field, and auditing is no exception. By embracing AI now, you’re not just keeping up with the times, you’re setting yourself up as a forward-thinking professional ready for whatever the future holds.

Far from replacing us, AI is here to make our auditing roles more engaging and fulfilling. AI is not just a tool, it’s our partner in crunching numbers and cracking cases. At AuditBoard we believe it is time to embrace AI and take our auditing game to the next level. Plus, when robots do take over the world, you’ll be on their good side!

P.S. I had AI advise me on how to maintain a consistent, fun, and engaging tone for this post. Let me know how it did!