Episode 15: Breaking Boundaries: How AI is Transforming Accounting — Lessons from Lem Chin Kok

The latest episode of the Chartered Accountants Global Update takes us on a fascinating journey into the intersection of technology, forensic accounting, and innovation. In this week’s episode, host [insert host name if relevant] shines a spotlight on Lem Chin Kok, whose career path and insights are redefining what it means to be a modern accountant.

From Police Officer to AI Pioneer

Lem’s story is anything but typical. He began his career as a police officer in Singapore’s Commercial Crime Investigation squad, tackling fraud and forgery long before data analytics became part of the accounting toolkit. From there, he spent 23 years building KPMG Singapore’s forensic practice before stepping into his current role as CEO of an AI-first technology firm.

His journey—from manual typewriters to cutting-edge artificial intelligence—embodies the rapid transformation our profession has experienced over the past three decades.

The AI Breakthrough That Changed Everything

One of the episode’s most striking moments comes when Lem recounts a breakthrough during his time at KPMG. His team was struggling with the limitations of early AI models that required “labeled data” to detect fraud. Then, a newly joined colleague suggested trying an unsupervised machine learning model on millions of transactions.

The results were astonishing: out of tens of millions of entries, the model identified ten unusual transactions, nine of which were confirmed frauds already charged in court.

This moment was a turning point—not just for Lem, but for the way AI could be used to uncover financial misconduct. The success wasn’t just about the algorithm; it came from combining forensic accounting expertise with data science. As Lem explains, the real power of AI lies in being guided by human understanding of accounting systems, fraud behavior, and judgment.

Why Accountants Should Learn to “Speak Machine”

Lem’s perspective on AI is refreshingly balanced. He divides it into two broad categories:

• Predictive AI – tools that can detect anomalies or risks in ways no human could replicate, such as identifying patterns across hundreds of data dimensions.
• Generative AI – tools like ChatGPT that enhance productivity by handling repetitive and manual work, freeing accountants to focus on strategic, high-value tasks.

But perhaps Lem’s boldest idea is that accountants need to learn programming languages—not to become coders, but to communicate effectively with technology. “Programming is a language,” he says, “just like English or Mandarin.” It’s the key to harnessing AI in meaningful, profession-specific ways.

Building the Future of Forensic Accounting

Lem has also been a driving force behind Singapore’s Financial Forensic Accounting Qualification, the first applied learning program of its kind in the Asia-Pacific region. The qualification blends forensic accounting, digital forensics, and financial crime investigation, equipping accountants to tackle the evolving risks of a digital economy.

The “AI Twin”: A Glimpse of Tomorrow’s Workplace

Now, as CEO of Arts, Lem is developing an extraordinary concept: the AI twin. This virtual twin can replicate the workflows of entire teams—performing tasks, following policies, making cognitive decisions, and producing outputs.

For audit firms, the implications are huge. Imagine AI handling the administrative grind while professionals focus on judgment, relationships, and strategic value. Lem sees this as the next frontier—AI not just answering questions, but executing full workflows.

Key Takeaways

1. AI won’t replace accountants—but accountants who use AI will replace those who don’t.
2. Accounting knowledge + technology skills = the new professional advantage.
3. We’re

Transcript

Welcome to Episode 15 of the Chartered Accountants Global Update, your weekly insight into the trends and innovations shaping our profession worldwide. Today we're doing something a bit different. We're taking a deep dive into the latest Difference Makers Discuss episode, featuring an absolutely fascinating conversation with Lem Chin Kok from the Institute of Singapore Chartered Accountants.

Now, I have to say, Lem's career trajectory is unlike any Chartered Accountant I've come across. We're talking about someone who started as a police officer responding to 999 calls, moved into white-collar crime investigation, spent 23 years building KPMG Singapore's forensic practice from the ground up, and is now CEO of an AI-first technology firm.

But what really captivated me about this interview wasn't just his unique background—it was his insights into how AI is genuinely transforming what we do as accountants. So let's unpack some of the key takeaways.

First, let me set the scene of where Lem started. Picture this: 30 years ago, a young police officer in Singapore's Commercial Crime Investigation squad, dealing with fraud cases, forgery, criminal breach of trust—and typing up all his reports on an actual typewriter. Fast forward to today, and he's at the cutting edge of AI technology specifically designed for the accounting profession.

That journey in itself tells us something important: the pace of technological change in our field has been absolutely exponential. But here's what struck me most—Lem saw this coming way before most of us did.

About 15 years ago, Lem had a breakthrough moment that completely changed his understanding of what AI could do for forensic accounting. He and his team at KPMG were trying to identify fraudulent transactions within massive datasets—we're talking tens of millions of transactions. They'd been experimenting with AI models but kept running into the same problem: most AI models need what's called "labelled data." Essentially, you have to teach the model what fraud looks like by showing it examples. But as Lem points out, the same fraud doesn't happen twice, so getting good training data is nearly impossible.

Then a junior colleague—someone who had literally just joined the firm—came to him with a different approach. She wanted to try an unsupervised machine learning model. Lem gave her team a huge dataset to test it on. When they came back, they showed him the ten most unusual transactions out of tens of millions.

And here's the kicker: nine out of those ten transactions were actual fraud cases that had already been charged in court.

Let that sink in for a moment. Out of tens of millions of transactions, the model identified ten highly unusual ones, and nine were genuine frauds. Lem said he almost fell off his chair.

Now, what made this work? This is crucial for us to understand. It wasn't just the AI model doing magic on its own. The secret was combining forensic accounting expertise with data science.

Lem explained that his team understood how accounting books and records function. They understood how fraudulent transactions behave. And they used that knowledge to build the features within the model. The model was operating in something like 20 or 30 dimensions—way beyond what human minds can visualise or process. But it was guided by accounting and forensic expertise.

This brings us to a really important point about AI and the accounting profession. There's been a lot of rhetoric—some people saying AI will make accountants redundant, others saying it's all hype. Lem's view is more nuanced, and I think he's spot on.

He breaks AI down into two categories for our profession. First, there's classical predictive AI—the kind he used to identify those fraud cases. This type of AI helps us do things that are humanly impossible, like analysing data in multiple dimensions simultaneously to identify anomalies we'd never spot through traditional sampling methods.

Then there's generative AI—the ChatGPT-style technology everyone's talking about. Lem sees this primarily as a productivity tool. It can take away manual and mundane tasks, freeing accountants and auditors to focus on higher-value work.

But here's what I found most interesting: Lem doesn't see AI as replacing accountants at all. In fact, he argues the opposite. He made his entire forensic team at KPMG learn programming languages. Not because he wanted them to become programmers, but because he sees programming as a language—no different from English or Mandarin—that allows accountants to communicate with machines and build models specific to their clients' needs.

Think about that for a second. We've always called accountancy "the language of business." Lem is suggesting we need to add another language to our toolkit—a programming language that lets us harness technology to do our jobs better.

And this isn't just theory. Lem was instrumental in developing Singapore's Financial Forensic Accounting Qualification—the first applied learning qualification of its kind in the Asia-Pacific region. It covers forensic accounting, digital forensics, and financial crime, and they're now partnering with law enforcement agencies across multiple countries.

Now, before we wrap up, I want to touch on what Lem is building now with his company, Arts. He's developed something called an "AI twin"—and this is genuinely cutting-edge stuff.

He describes it as a virtual twin of an entire group of employees that can perform all their tasks, following the organization's specific policies and procedures, with cognitive decision-making capabilities. It's not just about answering questions or automating repetitive tasks—it's about AI performing complete workflows, making decisions, and producing outputs, while humans focus on strategic, high-value work.

For audit firms, the implications are enormous. Imagine AI handling the time-consuming administrative and routine analytical tasks, while auditors focus on judgment, strategy, and client relationships.

So what are the takeaways from this conversation?

First, AI isn't coming for our jobs—but accountants who understand how to work with AI will have a significant advantage over those who don't.

Second, the combination of accounting expertise and technology skills is becoming essential. We need to start thinking about programming languages as tools in our professional toolkit.

Third, we're moving from an era of using AI to answer questions to an era where AI can perform complete tasks and workflows. That's a fundamental shift.

And finally, all of this reinforces that accountants remain central to the process. The AI that works best is built on accounting knowledge, guided by forensic understanding, and applied with professional judgment.

Lem Chin Kok is a remarkable example of how diverse experience—from law enforcement to technology to accounting—can position someone to see the future more clearly than most. I'd highly recommend listening to the full Difference Makers Discuss episode if you want to hear more about his journey and insights.

That's all for this episode of Chartered Accountants Global Update. Until next time, stay curious, stay informed, and keep pushing the boundaries of what our profession can achieve.