Latest / Difference Makers Podcast / Meet Lem Chin Kok - a cop turned accountant using AI to outrun fraud

Transcript

Meet Lem Chin Kok - a cop turned accountant using AI to outrun fraud · Difference Makers Podcast

• 0:15
  SPEAKER_00: Good morning, good afternoon, and good evening to our listeners across the globe.
• 0:20
  SPEAKER_00: And welcome to Chartered Accountants Worldwide Difference Makers Discuss.
• 0:25
  SPEAKER_00: I'm delighted today to be joined by our member from the Institute of Chartered Accountants Singapore, Lem Chin Kok.
• 0:33
  SPEAKER_00: I'm going to let Lem do his own introduction, but we are talking to Lem today on the basis that he has had a career in the police force, in cybersecurity, in forensics, and now is CEO of an AI first technology firm based in Singapore.
• 0:51
  SPEAKER_00: So Lem is going to talk to us about all those things that I think we know we need to be aware of and involved in cyber, AI, forensics, but maybe we don't know as much as we should.
• 1:03
  SPEAKER_00: So Lem, good uh afternoon to you.
• 1:06
  SPEAKER_00: And um Goodbye to you.
• 1:09
  SPEAKER_00: Thank you very much.
• 1:10
  SPEAKER_00: I have not done you justice um with the introduction.
• 1:12
  SPEAKER_00: So please give us a little bit of a plotted history from your start.
• 1:16
  SPEAKER_00: You started off in the police force and to where you are today.
• 1:20
  SPEAKER_01: Thank you, thank you.
• 1:21
  SPEAKER_01: So Sanit, uh, thanks for having me here.
• 1:24
  SPEAKER_01: Um so so yeah, I I started as a police officer some 30 years ago, or more than 30 years ago now.
• 1:31
  SPEAKER_01: Uh, and uh you you probably don't believe it when you dial triple nine, you know, I'll be in the fast response car uh to the crime scene.
• 1:41
  SPEAKER_01: Uh that's that's actually my first posting.
• 1:44
  SPEAKER_01: Uh and subsequently, I got posted into the commercial crime investigation squad uh where I do all the white-collar crime investigation.
• 1:55
  SPEAKER_01: Uh so I was there for uh five to six years, I believe.
• 1:59
  SPEAKER_01: And uh subsequently, you know, then uh when KPMG uh in Singapore uh started out uh with uh their forensic team, uh, I then joined uh the forensic team.
• 2:14
  SPEAKER_01: Um and and I've been there since.
• 2:17
  SPEAKER_01: Um and I just retired uh from KPMG after a good 23 years.
• 2:23
  SPEAKER_01: Uh and guess what?
• 2:25
  SPEAKER_01: Um, you know, uh when I retired, we have a team of uh about 100 strong uh forensic professionals.
• 2:33
  SPEAKER_01: Um and and it may sound a small number to you, but in a very, very small country like Singapore, uh where the crime rate is is low, it's actually quite a fairly big team uh that we have here.
• 2:46
  SPEAKER_01: Uh and yeah, and I'm now the CEO of Arts uh that I founded.
• 2:51
  SPEAKER_01: Uh and we are operating since a good uh eight months now.
• 2:57
  SPEAKER_01: Yeah, so so uh this is my background.
• 3:01
  SPEAKER_00: Wow.
• 3:01
  SPEAKER_00: Okay, so there's hang on, there's a huge amount here to unpack.
• 3:04
  SPEAKER_00: I want to go back to the fast moving car with the blue lights and all that.
• 3:08
  SPEAKER_00: Um there's not many accountants I talked to that have started their career as a cop.
• 3:12
  SPEAKER_00: Yeah.
• 3:12
  SPEAKER_00: Um, did you wait, did you start to train as an accountant when you were in the police force or which came first?
• 3:19
  SPEAKER_01: Actually, I I I started the interest before.
• 3:24
  SPEAKER_01: And and basically, you know, those were the years where we we really have to earn and we have to study.
• 3:32
  SPEAKER_01: Uh, and and as such, uh I signed on uh in the police force as a full-time uh professionals uh in in in the police force.
• 3:41
  SPEAKER_01: And during that time, you know, I I took up part-time studies again to to uh further um um uh prepare myself for for the world of accountancy.
• 3:56
  SPEAKER_01: Yeah.
• 3:57
  SPEAKER_00: Fantastic.
• 3:58
  SPEAKER_00: And and uh and okay, so you you went into as you called the the um white collar to investigate white-collar crime.
• 4:06
  SPEAKER_00: Um tell tell us a little bit about that.
• 4:08
  SPEAKER_00: Is that very much behind the desk kind of port through not no, okay.
• 4:12
  SPEAKER_01: Yeah, police force is never behind the stairs.
• 4:15
  SPEAKER_01: So so we have to do uh so so for example, cheating cases, um, there's uh criminal breach of trust cases, uh forgery cases.
• 4:26
  SPEAKER_01: Uh so we we have to be at the scene, we have to be talking to witnesses, uh, we have to be talking to uh the accused person, uh, we have to you know charge the accused person in court and we've got to appear in court.
• 4:41
  SPEAKER_01: Yeah, so so it's um it's it's very interesting.
• 4:44
  SPEAKER_01: And back then, guess what?
• 4:46
  SPEAKER_01: We we were using typewriter.
• 4:47
  SPEAKER_01: I I don't know uh a lot of the listener now probably don't know what how a typewriter looks like.
• 4:53
  SPEAKER_01: Yeah, but that's back then, you know, we were still using typewriter.
• 4:56
  SPEAKER_00: Really?
• 4:57
  SPEAKER_00: So you were collating all the kind of the records and all the rest on on typewriters.
• 5:01
  SPEAKER_00: Uh serious, seriously, um, seriously, Lem, you don't look old enough to to to to to be able to talk like that.
• 5:07
  SPEAKER_00: But okay, so you've obviously seen a huge progression in the use of technology.
• 5:14
  SPEAKER_00: And I mean, let's start with that, you know, going from typewriters.
• 5:19
  SPEAKER_00: Yeah, for loaded extremes.
• 5:21
  SPEAKER_00: What was, I mean, I'm not gonna ask the the question, what was the biggest jump that you've seen?
• 5:26
  SPEAKER_00: But but you know, I mean, you've obviously gone typewriters, computers, the boom in data, um, and then you went headlong into forensics.
• 5:35
  SPEAKER_00: Is there any kind of snippets you can bring out to to the audience that were the gifts of the yeah?
• 5:40
  SPEAKER_01: So basically, when I started uh my my career in KPMG in the forensic team, and back then, you know, it's tough because the team is uh it was really small back back in those days, and and we really need to differentiate ourselves.
• 5:55
  SPEAKER_01: And back then, in my mind, you know, the easiest to differentiate ourselves is really through uh the use of technology.
• 6:04
  SPEAKER_01: So we then bring in um digital forensic where we acquire uh evidence from uh an individual's computer, mobile phone, etc.
• 6:16
  SPEAKER_01: Uh, we then brought in e-discovery where we can ingest large amounts of uh uh data, and from there we see out uh the evidence that is critical for our investigation.
• 6:28
  SPEAKER_01: And uh we also go into uh forensic data analytics.
• 6:33
  SPEAKER_01: So that's uh you know our books and records, right?
• 6:36
  SPEAKER_01: So so it goes into millions or even billions of transactions and and how do we then use certain techniques to triangulate uh all these uh accounting books and records, uh all these data to identify you know potential problematic transactions, so to speak.
• 6:56
  SPEAKER_01: And actually the highlight uh you know to maybe draw some uh link to what I'm doing now.
• 7:03
  SPEAKER_01: So I'm very passionate with uh analytics, so I've been pushing for you know the the analytics um uh processes um I think as far back as 18 years.
• 7:18
  SPEAKER_01: Uh and and at about 15 years, that was the first time that I experienced AI.
• 7:27
  SPEAKER_01: And and you you you wouldn't believe uh what I've gone through.
• 7:31
  SPEAKER_01: So so I've been I I even hired professors from a university into the practice to to to um uh find a new way uh to detect fraud.
• 7:46
  SPEAKER_01: And and I was hoping that if I can use some uh AI technology to identify fraud from huge amounts of uh data sets.
• 7:55
  SPEAKER_01: And what happened was uh we tried different ways, and one day um uh a very junior colleague of mine, in fact, in fact, she just joined us, uh, if I if I recall correctly.
• 8:09
  SPEAKER_01: And she joined us and and she saw me trying different ways, uh putting in a lot of time to find new ways uh to identify problematic transactions from huge uh volume of records.
• 8:23
  SPEAKER_01: And he you know got his courage and came to my room and said, Hey, you know, Lam, uh, can you let me try something?
• 8:30
  SPEAKER_01: I I have this technique that I want to try.
• 8:33
  SPEAKER_01: Uh, I say, Are you sure?
• 8:35
  SPEAKER_01: Because we tried um uh you know AI models, but in a lot of instances, you need labeled data.
• 8:43
  SPEAKER_01: That means you need to know uh what is fraudulent, what is not fraudulent, and then the the the data the model will then train uh so that they can help to detect the next uh problematic data, uh problematic transactions, I mean.
• 8:57
  SPEAKER_01: Uh and and because in the world of fraud, uh the same fraud don't happen twice.
• 9:02
  SPEAKER_01: So so then you know it's it's it's a kind of uh scenario where you know it's very difficult to get labeled data.
• 9:11
  SPEAKER_01: But nonetheless, I say, okay, since you are you want to try, you know, uh you pull together a team, I I then give the team a very huge data set.
• 9:21
  SPEAKER_01: And after a while, they came back and they showed me, oh, Lam, these are the 10 most unusual transactions out of the tens of millions of transactions.
• 9:33
  SPEAKER_01: And when I look at it, I was like, I almost fell off my chair, but you know, I hold myself and I asked her, I said, why are all these 10 transactions unusual?
• 9:44
  SPEAKER_01: And she was looking at me, she's like, Um, Lamb, you know, we are using um multiple dimensions of analysis, and there are the trillions of possibilities.
• 9:57
  SPEAKER_01: There's no way that you will know.
• 9:59
  SPEAKER_01: Uh so so so, in a layman's term, is that you know, two dimensions is a picture, three dimension is an object, and can you visualize four dimensions, five dimensions?
• 10:10
  SPEAKER_01: But all these models go into 10, 20 dimensions, and and which is why it's so powerful, right?
• 10:17
  SPEAKER_01: Uh something that humans can't comprehend, which yeah, we can't.
• 10:22
  SPEAKER_01: Uh out of the top ten, I knew that nine of them were actually real fraud cases charged in court.
• 10:34
  SPEAKER_01: Not what I said, but charged in court.
• 10:37
  SPEAKER_01: Uh okay, so so the the the so you have tens of millions of transactions.
• 10:43
  SPEAKER_01: The model that the team built picked up ten highly unusual transactions.
• 10:51
  SPEAKER_01: And of the ten highly unusual transactions, nine were them were actually fraud cases that were charged in court.
• 11:01
  SPEAKER_01: Yeah, so so so from then on I was sold.
• 11:05
  SPEAKER_01: So I have been pushing uh AI since then, but as you know, um most of the time, you know, the individual sitting opposite me will give me a dirty look.
• 11:19
  SPEAKER_01: I I typically leave the room feeling very lousy uh back then.
• 11:23
  SPEAKER_01: Uh but yeah, nobody knows what I'm talking about.
• 11:26
  SPEAKER_01: Yeah.
• 11:27
  SPEAKER_01: Probably they thought I'm a con man.
• 11:30
  SPEAKER_00: But you know, I I mean uh evidently evidently you were ever able to kind of see into the into the the future before before the rest could.
• 11:39
  SPEAKER_00: Kimir, the what I find fascinating about that story is a couple of things.
• 11:43
  SPEAKER_00: So if I understand it correctly, and this is where sometimes different brains take longer to compute, it was a combination of skill sets that essentially came up with that finding.
• 11:57
  SPEAKER_00: So it was it was potentially an accountant who understood what fraud looked like.
• 12:01
  SPEAKER_00: It was there was maybe a technology expert, there was uh, you know, I and before I go on and and say something wrong, was it a combination of different skill sets?
• 12:12
  SPEAKER_00: And where does the accountant sit in that skill set formula?
• 12:17
  SPEAKER_01: Actually, you are spot on.
• 12:20
  SPEAKER_01: Yeah, you're actually spot on.
• 12:21
  SPEAKER_01: Yeah.
• 12:22
  SPEAKER_01: So so what happened is this uh what the the model, the technique that the team used at that time, uh is unsupervised uh machine learning model.
• 12:32
  SPEAKER_01: So so it's unsupervised.
• 12:34
  SPEAKER_01: So what it then meant is that you don't need labeled data.
• 12:37
  SPEAKER_01: So basically, a huge amount of data, the model will learn by itself.
• 12:42
  SPEAKER_01: And the trick is that you know, because for us we understand accounting books and records, how it functions.
• 12:50
  SPEAKER_01: And we also, because as forensic professionals, we also understand how a fraudulent transaction will behave.
• 12:59
  SPEAKER_01: And and with these skill sets, the you know, the our understanding of how books and records were maintained, and uh the underlying characteristic of fraud, we actually use this um uh knowledge to build the features in the model.
• 13:19
  SPEAKER_01: So so you know, predictive AI models is basically controlled by uh the features in there.
• 13:26
  SPEAKER_01: Uh and and which is why there is unlimited permutation of how you can build the model.
• 13:33
  SPEAKER_01: And and basically, when we were sharpening it with uh you know our accountancy knowledge, uh bookkeeping knowledge, our knowledge of uh how the the fraud characteristic, uh it was so accurate.
• 13:49
  SPEAKER_01: So so in short, uh we can have another data scientist uh with the same data set and try to run some models uh to identify fraud within this model.
• 14:03
  SPEAKER_01: Uh, very likely uh you know the model won't perform as good as uh what uh my team at that time uh performed.
• 14:13
  SPEAKER_01: Yeah.
• 14:15
  SPEAKER_00: Okay, so it you know, there's a lot of noise about AI and the profession.
• 14:21
  SPEAKER_00: And there's probably there's two extreme views, and then there's a lot of views in the middle.
• 14:25
  SPEAKER_00: And certainly in Ireland, there's been a rhetoric in the last few months where um people have come out and said AI is going to make the accountant redundant in the future, and you know, it's gonna take all our jobs, and and then you've got the uh others who are saying, well, look, that's absolute nonsense because A, the accountants are needed to feed into the AI development, as you've just talked about, but B, what will happen is it'll reduce all the kind of manual tasks and therefore let the accountants get on with doing the bigger things.
• 14:56
  SPEAKER_00: I think I know where your view is, but how big is the opportunity of AI in the profession?
• 15:05
  SPEAKER_01: Uh but to answer that question, I think is somewhat in between.
• 15:09
  SPEAKER_01: Okay, I think it's somewhat in between.
• 15:11
  SPEAKER_01: Um so the the the whole value of AI that I see today uh is twofold.
• 15:20
  SPEAKER_01: Uh one is that uh because there are many kinds of AI models.
• 15:25
  SPEAKER_01: So the examples that I gave you just now is a predictive model, the classical predictive model uh that has been around for the longest time, you know, probably older than me.
• 15:37
  SPEAKER_01: Uh and and what we have today that that you you you know that you saw, you know, the Chat GPT, the DeepSeek.
• 15:47
  SPEAKER_01: Uh this is uh what we call the generative AI.
• 15:51
  SPEAKER_01: So so it's it's actually two two different uh two very distinct and different types of uh AI models.
• 15:59
  SPEAKER_01: Um so so basically the classical AI models will enable auditor, accountant to predict, to identify the transactions that we need to identify in a way that is humanly impossible.
• 16:20
  SPEAKER_01: Like, you know, the example just now we would have gone into high-dimensional space.
• 16:25
  SPEAKER_01: The model could be operating in 20 dimensions, 30 dimensions, that it's just humanly not possible.
• 16:32
  SPEAKER_01: And and in the world today is actually quite important because our our operating environment involves a huge volume of transactions, right?
• 16:41
  SPEAKER_01: So so basically uh you need as accountant, I I strongly uh recommend that you you need to get into the technology so that out of the millions of transactions that you audit, you know, it can help you identify certain transactions that you need to identify for your audit purposes, for example.
• 17:07
  SPEAKER_01: Whereas on the other side, the generative AI space is really a productivity play.
• 17:15
  SPEAKER_01: And basically it can, like what they say, uh a huge potential there is that it can help you take out the manual and mandane tasks so that uh accountant, auditor can perform higher value tasks.
• 17:33
  SPEAKER_01: Because we we also don't want to be bogged down by all the manual and banding tasks, right?
• 17:37
  SPEAKER_01: So so the the the generative AI models uh that is uh becoming very, very popular these days, yeah, they they actually have a real uh uh benefit of helping to take out all the manual and mandate tasks so that the the accountant can perform the real job and the valuable job.
• 18:02
  SPEAKER_01: So so um uh I think this will be central to our industry.
• 18:08
  SPEAKER_01: One is to make us more effective uh in in the current world, because I seriously don't believe that by picking samples randomly it can help us a lot when you are dealing with uh millions and millions of transactions.
• 18:25
  SPEAKER_01: So you need certain techniques to help us to identify the transactions that we want to identify.
• 18:32
  SPEAKER_01: And on the other hand, because you're dealing with millions and millions of transactions, and the kind of uh indirect administrative work on the auditor then becomes uh tremendous.
• 18:46
  SPEAKER_01: So so you also want to use uh certain technology to help you take away the manual and maintaining tasks so that you know we can focus on doing the important tasks.
• 19:00
  SPEAKER_00: Thank you, Lem.
• 19:02
  SPEAKER_00: What you have just educated me in in the last four minutes is probably something that I've struggled with for the last 10 years.
• 19:08
  SPEAKER_00: You've explained it in a way that that is made very simple to me, uh, and I appreciate that.
• 19:14
  SPEAKER_00: Um in that regard, uh I'm passionate about the next gen and what they need to be armed with to be a success in that in the profession.
• 19:24
  SPEAKER_00: In that vein, do you feel we as a profession should be educating our accountants of today and tomorrow more in I want to call it technology, I want to call it programming.
• 19:39
  SPEAKER_00: I I don't know what what what we call it, but what's your view?
• 19:43
  SPEAKER_01: Okay, so um you know what I make my entire team in the past when I was still in in KPMG, uh I actually encourage everyone within the team to take up a third language, which is programming language.
• 20:01
  SPEAKER_01: So so you see, all these models are actually developed with uh certain programming language, and to be honest, all this language, if I look at it uh in in a very objective way, it is no difference from English language, Chinese language, you know, Japanese language.
• 20:23
  SPEAKER_01: So basically, uh the knowledge of an additional uh language will allow uh accountants, auditors to uh use the language uh to get what we want uh and to achieve what we want to achieve.
• 20:44
  SPEAKER_01: Um so so very similar uh in you know we we have English knowledge and we are able to communicate, we are able to write reports, and likewise, if we are armed with programming language, uh looking at the situation with the data set of this particular client, with the risks of this this particular client, the processes of this particular client, I can then use my programming language to build certain model that is uh specific for this client.
• 21:17
  SPEAKER_01: And when I apply the model, it it can then help me as an accountant or auditor to detect transactions that I want to detect out of their millions and millions of transactions.
• 21:30
  SPEAKER_01: So so um I have been talking about it uh with my peers, with my friends, uh, that you know, I think we are in an era that um we need knowledge of a third language or a second language.
• 21:46
  SPEAKER_01: Um yeah, I actually it's not just for accountants, it's uh I look at it as in for all industry.
• 21:54
  SPEAKER_00: Okay, okay.
• 21:55
  SPEAKER_00: That's fascinating.
• 21:56
  SPEAKER_00: And and and as you were talking, I mean, accounting has been defined as the language of business or the international language of business.
• 22:04
  SPEAKER_00: But what you're kind of saying is if you can layer on the the uh the tech tech language or the data language, then then it opens up so much more.
• 22:11
  SPEAKER_00: Um, I think that's amazing.
• 22:13
  SPEAKER_00: Um, there's a couple of things I want to touch on.
• 22:15
  SPEAKER_00: Um, I think to listeners, it would be no surprise the way that you have taught us in the last few minutes and articulated to hear that you were the founder of the forensic diploma.
• 22:28
  SPEAKER_00: Is that is that right in the institute?
• 22:30
  SPEAKER_01: Yeah, it's right.
• 22:31
  SPEAKER_01: Yeah, yeah.
• 22:31
  SPEAKER_01: So so I think it was about eight years ago.
• 22:34
  SPEAKER_01: Uh together with uh uh Iskar, we actually developed uh the financial forensic accounting qualification and we launched it uh formally about six years ago, if if I recall correctly.
• 22:49
  SPEAKER_01: And it is actually the first applied learning financial forensic qualification uh developed by a professional body in in this Asia-Pacific region.
• 22:59
  SPEAKER_01: Uh so it's uh uh we actually focus on on three key areas, uh the forensic accounting and investigation.
• 23:08
  SPEAKER_01: Uh the second area is on digital forensic.
• 23:12
  SPEAKER_01: So, you know, very much uh what I shared earlier, uh to get evidence from uh computers, mobile phones, to get evidence from large data sets, to apply models, algorithms, you know, analytics to identify anomalies.
• 23:30
  SPEAKER_01: Uh so uh that that that kind of uh skills.
• 23:35
  SPEAKER_01: And uh what's interesting is that we also have uh the third part, which is on financial crime, which is um uh in in the financial services sector is uh a lot of focus on uh money laundering, investigation, sanctions, investigations, uh misconduct, scams, etc.
• 23:59
  SPEAKER_01: So it's uh it's it's a qualification that yeah very, very broad and and looks across uh a wide spectrum of how accountancy together with forensic practices can be applied in the industry.
• 24:17
  SPEAKER_01: Yeah.
• 24:19
  SPEAKER_00: That's that's that's phenomenal.
• 24:20
  SPEAKER_00: And is it is that something that that accountants outside of Singapore can can sign up for?
• 24:26
  SPEAKER_00: Because as you said, your country is very is a very clean country, but there's a lot of countries that could produce this.
• 24:32
  SPEAKER_01: Yeah, so so um I still recall, you know, um probably a year back or two years back, so so we started uh to to go into partnership with uh the establishment, like for example, the corruption uh practices and investigation bureau uh in in Singapore, the Singapore Police Force, and also for financial intelligence and investigation bureau of the Hong Kong Police Force.
• 24:59
  SPEAKER_01: So I think um yeah, we we we back then when we I'm still the chair of uh the oversight committee, uh we have been trying to uh sign partnership uh with not just entities within Singapore, but also outside of Singapore.
• 25:17
  SPEAKER_00: Wow, congratulations, Lar.
• 25:19
  SPEAKER_00: Um this is fascinating.
• 25:21
  SPEAKER_00: I could talk to you forever, but we're we're we're we're coming up to the end of time.
• 25:25
  SPEAKER_00: We may have to do a part two because I really do I know that this topic is is really um you know um key to members.
• 25:32
  SPEAKER_00: And I know Charge Countons Worldwide did a survey, and AI was considered one of the, I suppose, one of the most crucial um issues uh facing the profession at the moment.
• 25:41
  SPEAKER_00: But before we finish, you know, I do want to talk a little bit about arts.
• 25:45
  SPEAKER_00: Um, you know, you your CEO of the company, you're you're you're you're a founder.
• 25:50
  SPEAKER_00: Um tell us what your aims or ambitions or you know vision is for that.
• 25:57
  SPEAKER_01: Yeah, so so uh for us really is uh to help our client to use AI in a way that is not used before.
• 26:11
  SPEAKER_01: So currently, if you look at AI, you'll think about you know the Chat GPT, the the Deep Sick.
• 26:18
  SPEAKER_01: Uh but in the enterprise world, actually, there are many other creative ways to use uh AI.
• 26:28
  SPEAKER_01: Uh and and what we really want to do is to help company enterprises derive significant real-world benefit from AI.
• 26:43
  SPEAKER_01: Um, as of now, um, maybe you're aware as well, AI is very much skewed towards the consumer market.
• 26:51
  SPEAKER_01: So, as consumer, you know, my family members love it because uh it is is is it's quite useful.
• 26:57
  SPEAKER_01: But the randomness of the technology when you try to apply it for an enterprise is really very difficult.
• 27:06
  SPEAKER_01: Uh, I don't know whether you're aware, if if you have uh you and your colleagues, maybe three or four of you, you have the same prompt and you try to press enter key together uh in the same uh application, the answers that come back uh is all different.
• 27:24
  SPEAKER_01: So so what it then meant is that you know that the technology is actually by nature very random.
• 27:31
  SPEAKER_01: And in the enterprise world where you have your policies, your procedures, uh the last thing you want is that you, you know, your 5,000 staff doing whatever they they want, right?
• 27:42
  SPEAKER_01: Yeah, so so um what arts is is um in fact we have a patented technology uh and we trademarked it as well.
• 27:51
  SPEAKER_01: It's called the Gen AI twin.
• 27:56
  SPEAKER_01: Uh so so basically, this is a technology that is um using AI in a way that is not seen before.
• 28:09
  SPEAKER_01: So, what can it do?
• 28:10
  SPEAKER_01: Maybe as uh my my parting message to our audience here.
• 28:16
  SPEAKER_01: Basically, Gen Eye Twin is like a virtual twin of a big group of your employees, a big group, not one, uh, in any roles.
• 28:29
  SPEAKER_01: It can be in finance, it can be in sales, it can be in compliance, it can be in operations, etc.
• 28:36
  SPEAKER_01: In any roles.
• 28:38
  SPEAKER_01: The Gen AI Twin perform not minor tasks, but Gen AI Twin perform all the tasks that this group of employees perform from morning to the night to produce output that this group of employees produce.
• 29:00
  SPEAKER_01: So the Gen EI twin will produce the same output and uh preferably with better quality.
• 29:08
  SPEAKER_01: And and the most important thing is that as the Gen AI twin performed the tasks, uh, it performed the tasks per the unique policy and procedures of that organization.
• 29:21
  SPEAKER_01: So so um what we're trying to do is for the first time, uh it's not a situation where you use uh AI, you query, you prompt them to get an answer, and you still need to do the work.
• 29:36
  SPEAKER_01: Uh in arts, we want our AI technology to perform the tasks itself so that you know uh we can help take out the manual and maintain tasks, and then the valuable resources can then be having more time to perform higher value tasks.
• 29:59
  SPEAKER_00: Okay.
• 30:00
  SPEAKER_00: I I I stayed with you for probably about 90% of that.
• 30:04
  SPEAKER_00: And I got got a bit lost.
• 30:06
  SPEAKER_00: But I mean to my head, that sounds like it's it's it's it's not just the manual repetitive tasks.
• 30:15
  SPEAKER_00: It is now getting into the space of group tasks and and potentially judgmental areas.
• 30:22
  SPEAKER_01: Yes, of course.
• 30:23
  SPEAKER_01: Of course.
• 30:24
  SPEAKER_01: So so basically you I mean in our workplace, we need to hire this group of people because we need their cognitive capability to make decisions for the organization from morning to the night.
• 30:36
  SPEAKER_01: So the Gen AI twin will be perform will be performing the task of this group of people because the Gen AI twin will now have the cognitive capability and to perform the tasks with uh and make cognitive decisions from morning to night for the organization, for the organization's policy and procedures.
• 31:00
  SPEAKER_00: Wow.
• 31:01
  SPEAKER_00: Wow.
• 31:01
  SPEAKER_00: Lem, the the I'm beginning to kind of see, well I can't see, but the endless possibilities.
• 31:08
  SPEAKER_00: And I mean I'm thinking I'm an I'm an I'm an ex-audit partner.
• 31:11
  SPEAKER_00: I'm kind of thinking the impact this could have on auditing is is is really phenomenal.
• 31:16
  SPEAKER_00: But but again, leaving the accountant the headspace and the ability to do other tasks.
• 31:23
  SPEAKER_00: So it's not it's not taking our jobs, it's just potentially also helping us do our jobs in a different way.
• 31:29
  SPEAKER_01: Better.
• 31:29
  SPEAKER_01: So so basically, you see, uh I I mean I used to be with the firm of you know three, four thousand people, and and a lot of time you you'll notice that uh a lot of their time are being sucked away to perform manual and mandate tasks uh where they are important, uh but may not be as crucial as the key tasks.
• 31:53
  SPEAKER_01: Uh and because they they were you know a lot of the resources is taken away, then they have very limited time to really focus on the key tasks.
• 32:03
  SPEAKER_01: So now what we are saying is that if if the technology can help to take away to perform uh the manual and mandate tasks, yes, they are important.
• 32:15
  SPEAKER_01: So so whatever that is performed as humans will still come in to check, yeah, it this is good.
• 32:20
  SPEAKER_01: Yeah, and they are making the right decision per the policies and procedures, uh per what the organization wants uh to be performed.
• 32:30
  SPEAKER_01: And then with all this uh time that is free up, they can then focus and do a proper audit uh and and do a proper strategy for for their client.
• 32:43
  SPEAKER_00: Yeah, yeah, yeah, I get it.
• 32:45
  SPEAKER_00: Um and I love it.
• 32:46
  SPEAKER_00: I love it.
• 32:47
  SPEAKER_00: It's it's it's it's really actually it's it sparks something in me.
• 32:50
  SPEAKER_00: And I think there's so much that the educators of to you know today need to take from this.
• 32:55
  SPEAKER_00: Lem, I'm gonna pull this to a close.
• 32:58
  SPEAKER_00: I think this has been fascinating, certainly for me, and I hope the listeners have enjoyed it as much as I have.
• 33:03
  SPEAKER_00: Um, I wanna I want to congratulate you.
• 33:06
  SPEAKER_00: I think you are an amazing ambassador for our profession.
• 33:10
  SPEAKER_00: I think you have that um mindset that people are gonna potentially uh think of when they think of accountants, and that's why I'd love this to be seen by by loads and loads and loads of people because you really are a great ambassador for the profession, and what you are developing and and and bringing to all facets of business is fantastic.
• 33:31
  SPEAKER_00: So thank you very much, Lem, for taking the time to speak to us today.
• 33:35
  SPEAKER_00: And um, I wish you all the best with arts in the future.
• 33:39
  SPEAKER_01: Thank you very much.
• 33:40
  SPEAKER_01: Thank you.
• 33:43
  SPEAKER_00: So thanks very much to Lem.
• 33:46
  SPEAKER_00: Um, and thank you to everybody um for listening to us today.
• 33:49
  SPEAKER_00: I hope you thoroughly enjoyed this.
• 33:51
  SPEAKER_00: Um stay tuned for the um the next episode of Difference Makers Discuss, where we will be um bumping around the globe um interviewing charge count members.
• 34:02
  SPEAKER_00: Thanks a million.
• 34:03
  SPEAKER_00: Bye.