Firmcheck: Artificial Intelligence (AI) and the human element in AML compliance

The “state” of AI

Artificial intelligence (AI) is certainly not going anywhere – its popularity continues to grow globally, you only have to read 5 recent headlines from various Gartner research reports to understand the seismic shift that AI is having globally:

• Gartner predicts AI software will grow to $297 billion by 2027

• Gartner survey finds 79% of corporate strategists see AI and analytics as critical to their success over the next two years

• Sixty-eight per cent of finance organisations are using AI or plan to use the technology,

• By 2026, more than 80% of enterprises will have used generative artificial intelligence

• Of more than 1,400 executive leaders, 45% reported that they are in piloting mode with generative AI

And if we then zoom in specifically on the accounting software industry the big players like Xero and Sage have announced significant AI-driven developments, and that’s barely scratching the surface. Both the Digital Accountancy Show (DAS) 2024, and Accountex London had significant parts of their seminar programmes loaded with content on using or adopting artificial intelligence in the accounting and bookkeeping space – all of which seemed to have the theatres brimming to the edges.

At DAS we even saw several keynote speeches from one of the frontrunners of AI in the accounting industry, Jason Staats. He captivated the entire show during his talks. People want to learn about it, but similarly, we’ve also seen a sense of fear of replacement, there have been some emerging pieces of research that demonstrate there is some worry among the accounting profession that AI may put them out of a job.

As with any generational transformation, there is a hype cycle that ensues, and right now we are still very much at the beginnings of truly understanding how AI can help us, but what we need to avoid doing is seeing it as an existential threat to the accounting profession –– one thing that accountants and bookkeepers will always have on their AI counterparts is that they are human, and in this hype cycle of AI we need to remember that we are a what truly makes the difference when it comes to improving the lives and businesses of our clients.

So for a moment let’s try to break away from the “AI robots are coming to replace us” headlines, and reflect on what will remain familiar, or our accounting firms may even benefit from.

A brief nod to the humble spreadsheet

Let’s look back at history – remember Excel, still one of the world’s most popular tools. I had a challenge finding exact statistics but my research suggests that we can expect somewhere close to 1 billion users still using Microsoft Excel.

Spreadsheets came around at a time of great digitisation in the world and at the time Excel was in some cases an existential threat to people who made a living from some good old-fashioned number crunching using pen, paper and calculator. However, we’re here, some 30 years later and still using the humble spreadsheet, in fact still truly trying to understand how to get the most out of them. Largely the jobs impacted by the advent of Microsoft Excel still exist, they just exist in a slightly different capacity as we’ve adapted to harness the technology.

And so whilst AI is a disruptive force, we’re barely at the beginning of our understanding of what it can do, we need to be pragmatic about what it won’t replace, and how we can adapt what we can do as accounting professionals, as opposed to running terrified that AI is going to take over our industry.

The human element of AML compliance

We know AI technology is excellent for repetitive tasks, those that involve large quantities of data, such as data entry, and those that are time intensive, which lucky for us in the accounting industry a large portion of the work we do behind the scenes has one of those elements.

It doesn't necessarily need to take over the task entirely either. Some smarts behind the scenes that make a specific task quicker due to pattern recognition for example is a good example of how it can be helpful, but there still is a need for a human touch.

In the context of AML compliance we already see AI being used for:

• Transaction monitoring and identifying suspicious activity

• ID verification

• Identifying high-risk individuals (PEPs, Sanctions & Adverse Media)

And over time we no doubt might see it become more involved in other parts of AML compliance such as risk assessments, but this is where we enter a grey area in the world of AML compliance, as whilst a “risk-based approach” seems like logic an AI can easily interpret, the downside is that risk is innate subjective to an individual. There are absolutely some instances where there might be hard rules around choosing not to work with a client, but there are some instances where you as the Director of a firm, are comfortable with the level of risk you’ve assessed, you’ve documented that, and you’re OK working with that client.

It’s also not uncommon for a large number of false flags to be identified, especially when running PEPs and Sanctions checks – common names in particular are synonymous for throwing up red flags. Whilst we might develop more robust AI to identify red flags in the future, a level of human involvement is still required to understand that risk, and what’s accurate or not.

And even with the help of AI, whatever it might flag as part of your onboarding or KYC process, you as a human still need to review whatever has been flagged and decide if it’s out of character or simply a mistake based on what you know about the client.

Take Biometric ID verification for example, a poor photo of a passport because of bad lighting, might cause the system to “fail” an identity check, but when you review the document you see that it was just a case of bad lightening and the technology couldn’t quite figure it out – if we left this work to the “robots” we might end up making this more of a burden than we want it to be.

So whilst a large % of us see AML compliance as a check box exercise that we don’t necessarily want to spend time doing, a critical part of what we need to do to meet our AML compliance obligations involves empathy and ethical judgment both of which are built on trust with clients.

AI in AML compliance should be viewed as an opportunity to enhance efficiency and accuracy rather than be seen as an existential threat to the accountants and bookkeepers involved. Leveraging AI to help in the manual administration of AML, and automating routine and data-intensive tasks is a no-brainer, so as humans, we can focus on the more complex and judgemental aspects of AML compliance.

Ensuring we’re AML compliant, and have a robust process built around risk-based decision-making is critical to avoid any hefty fines, and where possible we’re leveraging technology, and maybe even AI to augment what we do.

Preparing for the future with AI

So how do you prepare for a world where AI is more ingrained into your everyday work? Truth is you’re probably already there, but the simplest and easiest way is to level up your knowledge of where it can be helpful.

How I’ve been doing this in my role is looking at what’s consistent, versus what’s not.

Things like data reports from website traffic, or even internal team messages – essentially anything where there is a lot of data that used to require some level of diving into, I can now get a summary or punchier overview of what’s going on, more quickly. Another example is writing, this article started life as an outline generated by AI, I then wrote the article, and then used AI again to help me optimise headings and make for better reading – the key things there though are that I understand the audience, and I’ve guided the AI to the outputs it gave me, it just saved me a heap of time. Another example of the relationship between AI and Humans, it’s not all or nothing.

The time AI can create for you to spend more time on elements that need to be human-led is quite powerful, but the public narrative sometimes leans into the ‘fear’ angle of AI, which isn’t particularly helpful – we’re certainly not in an all-or-nothing situation with artificial intelligence.

Preparing for a future with AI in it is simple. Harness what it’s capable of right now, keep an eye on how it’s changing, and invest in what makes humans great, and that’s building trust and excellent relationships with your clients. All of those interactions you have with your clients are the ever-evolving building blocks of truly understanding your customer, with that being able to provide the best service possible, and also in a ‘hidden’ way building a picture that ultimately feeds into your AML compliance management, and ongoing risk analysis.

Don’t be ignorant of artificial intelligence, but also it’s not an existential threat to our industry. Humans have adapted time and time before, just like we did with the humble spreadsheet, and the industry will evolve similarly again, perhaps just more rapidly than the previous cycles of digital transformation.

(NB: This article doesn't constitute legal advice and is only intended for general informational purposes. Always consult with a legal expert or compliance consultant for guidance specific to your firm.)