According to certain estimates, regulators across the world handed out $8.4 billion in anti-money laundering fines in 2019. Global KYC solutions provider Encompass Corporation estimates that authorities handed out a record 58 anti-money laundering penalties across the world. This is a 100 percent increase over the 29 penalties worth $4.27 billion imposed in 2018.
Approximately, half of the companies that paid AML penalties last year were banks. Obviously, the losses for financial institutions that are hit with AML penalties are not limited to just the penalties themselves. They pay a heavy price in terms of declining revenues, customer dissatisfaction, collapsing stock prices, and loss of reputation and brand value.
Rules-based software that used to be deployed by banks to detect money laundering is inefficient and leaves analysts with too many alerts to deal with that could be positive or negative. Solutions based on artificial intelligence are much more efficient in detecting money laundering, compared to software-based on rule-based approaches.
Today deep neural networks can reveal complex interdependencies among money laundering activities across the world leading to fewer false alarms and more accurate recommendations. One such fintech startup that stands out for its artificial intelligence-based AML software is Silent Eight, which is also one of the many fintech companies founded in Singapore by European entrepreneurs.
In the case of Silent Eight’s solution, the recommendations are supported by a written narrative explaining in plain English the decisions. Silent Eight’s machine continually learns as time goes by, and updates its algorithms to constantly improve the quality of its recommendations. The result is that it significantly reduces analysts’ time to review cases and arrive at correct conclusions.
The rising number of global fintech entrepreneurs who are heading to Singapore to launch their fintech startups stands testimony to the allure of the city state’s world-class business environment, regulatory framework, and global mindset. Coming from Poland, Martin Markiewicz is the co-founder and CEO of Silent Eight. Based in Singapore, which is now a major regtech hub, Martin provides the vision behind the company’s AI-based advancements in fighting financial crimes.
Silent Eight is the winner of the FinTech Abu Dhabi Innovation Challenge and the Monetary Authority of Singapore’s 2017 Fintech Hackcelerator award. In 2018, it won a top fintech award in Australia. The same year, Standard Chartered announced that as part of its efforts to lead the way in the global fight against financial crime through the use of regtech, it had partnered with Silent Eight to deliver cutting edge capabilities to its Financial Crime Compliance (FCC) teams.
In 2019, Standard Chartered’s fintech and ventures unit, SC Ventures invested in Silent Eight’s oversubscribed Series A funding round, becoming a minority investor in Silent Eight and reaffirming the global banking giant’s trust in Silent Eight’s solution.
With an educational qualification in mathematics, Martin is a serial entrepreneur. Before launching Silent Eight, Martin had made his mark, creating a few successful startups in Europe and Asia, which includes a startup that saw a successful IPO.
In fact, Martin launched his first startup, Konsultant.it, which provided software and hardware development solutions for small and medium enterprises in 2001. In between, he was the strategic sales director of Wola Info, a leading European IT company. Later Martin established SevenFlow Investments – a multidisciplinary engineering company with a track record of landmark projects. SevenFlow Investments would finally become a part of a highly successful IPO.
With his 16 years of experience in software and artificial intelligence solutions covering a wide range of applications, Martin has taken the challenge of helping banks outsmart financial criminals and money launderers, who are gaming their transaction systems, headlong. In an exclusive interview with International Finance, Martin speaks about the Singapore fintech startup ecosystem, the value proposition that Silent Eight provides to its users, the Singapore fintech’s growth, and the future of AML software.
International Finance: Could you tell us more about the background of Silent Eight founders and the motivation to launch an anti-money laundering startup in Singapore?
Martin Markiewicz: Before we started Silent Eight, we took a company from startup to publicly traded in Poland. Our track record of creating a publicly-traded company from an idea gave us the confidence to try something new.
We were looking to do something significant, something that would help people. Our strengths are in engineering and problem solving, so we were looking for a global problem we could solve that would make the world better. It sounds a little cliched, but it’s what we wanted to do. We kept coming across money laundering, financial crime, the challenges that institutions face to run their business and ensure exactly who they were doing that business with.
In essence, we understood the global damage caused by all these activities. We saw the billions of dollars being spent every year to fight financial crime, and we also saw statistics after statistics that showed the bad guys won way too often.
The way we build is from the bottom up, working with a customer to solve a problem. This led us to Singapore, a global financial centre, to launch our business in supporting banks to combat global financial crime.
If you examine our clients, they are clients with a global scale and to match them, we are expanding globally with offices in Singapore, New York, Chicago, Seattle, London, Hyderabad, and Warsaw.
What are the key challenges that AML solutions face globally, especially the volume of false alerts? How do Silent Eight’s solutions minimise this challenge?
The false alerts are just one of the many problems. There are a lot of major lawbreakers and bad guys who are trying to get into the financial system and freely move around it. What financial institutions need to do is investigate existing and potential clients, vendors, and other partners in terms of their activities.
We help with false alerts.
We investigate 100 percent of the alerts for our customers and solve them. We do not suppress them nor are they weighted or partially solved; they are either solved or not. This is one of our key differentiators. Our IP does not decide whether an alert should be solved or not and which way it should be solved. Our AI acts according to the specifications of our clients.
The removal of false alerts is the first step to achieve our purpose, which is finding true alerts. Solving false alerts is challenging, and it’s important. Finding true positives is what we are here to do. We help our customers keep clear of terrorists, drug lords, and sanctioned people, and this way we also secure the interests of firms that protect our clients and the broader financial system. But ultimately, it’s even bigger than that in scope.
Each time we help a client refrain from financing a person with bad intent, we make it that much harder for the person with bad intent to hurt people at scale. That’s what we are about; solving false alerts gets you to true alerts and solving true alerts saves lives, money, and jobs. This is exactly what we were looking to do.
So, what are the implications of Standard Chartered’s investment in Silent Eight?
Standard Chartered is one of our minority shareholders. One of the implications is that they have invested in us and that gives us a vote of confidence. It’s always a good feeling when a client using your product says “hey, can I invest in your company?” I hope the implication of this is that we will do great in the future.
I feel that our product is differentiated, and it’s a great fillip for a global bank to invest with their capital in a company that provides them a critical product, not limit their commitment to just words.
With regard to regtech, ongoing developments like the US-China trade war are creating new sanctions lists. How is Silent Eight keeping up with these dynamic developments?
We do not create or maintain sanctions lists. Our customers have other excellent providers for that. Our AI leverages those lists and learns and grows each time an input changes, including a sanctions list. This approach means no matter what new sanctions are added or changed we are ready to support our customers in abiding by their directives.
In terms of minimising human effort and maximising productivity, can you quantify the gains that organisations can make by using Silent Eight solutions?
It’s clear that the AI can process data and solve alerts at a velocity unreachable by humans. However, we do not view it as a clear choice between human or AI solved alerts. We see it as a very traditional AI-human relationship in that a human sets the rules, the AI does the work, and another human checks the work. It is symbiotic with each component in the chain responsible for the part they are best suited.
The other key differentiator between AI and human solutions is the AI is incapable of making a mistake, either through poor training, or bias, or tiredness or any of the flaws that we are made of.
On a daily basis you may have thousands of alerts. So how does Silent Eight system ensure that it scales to meet the hundreds of thousands of alerts?
Our system is built under a scalable infrastructure with a capacity to handle hundreds of thousands of alerts every day. It’s designed to work with the biggest financial organisations in the world across multiple jurisdictions and languages.
With the number of machine learning based AML solutions available in the market, what is the unique value proposition offered by Silent Eight to financial institutions?
The first key differentiator is there is no opaqueness in how an alert is solved. We show clients each of the agent results that created the solution and which set of client rules it followed. And each alert is auditable.
It’s important to reiterate, we do not weight, or recommend, give probabilities, or suggest, and we definitely do not suppress. We follow the rules the clients give us. It is as simple as that.
Each time we solve an alert we tell the overseeing analyst exactly why the alert was solved without any black box challenges. We provide the solution and the explanation how it is worked out.
For example, it is harder for clients to trust a score generated by the machine stating that “this case has only 5 percent probability of something serious.” In my opinion, that approach doesn’t help very much. On the other hand, what helps is the machine saying “Hey this is a true case or a false case,” and then further explaining the recommendation with supporting information. With that, clients can agree or disagree with the recommendation and justify their final decision. The idea is to ensure that the machine generates transparent, reliable, and explainable information.
Which are your key markets and where do you see the most growth in the next five years?
We have big projects with European and US domiciled global banks underway and many regional banks especially in the US. We really see our target market as anyone who is worried about the accuracy and consistency of their AML processes and who could do with an AI-based helping hand.
What are the advantages and challenges of scaling a regtech startup in Singapore?
As I mentioned earlier, Singapore would be the best starting place for establishing a business. I think the fact that we started off in Singapore is an advantage because of its great regulatory and business environment.
Singapore offers a great platform for us to build something like this. Also, it is a pretty small market unlike the US — so the mindset is to go global right from the start. We are currently surrounded by like-minded people with a similar approach. That’s definitely an advantage.
In terms of developing technology, what are your plans for the next five years?
Our technology is receiving awards nearly every month, so we are very happy with the status quo. We, like any client-focused business, continue to develop based on client feedback and we have a full roadmap of client-driven requests.
Right now, we are on a certain version of the product, and the next version will be much, much better than the current one. We will ensure that the product only gets better across all use cases.
What does the future of machine learning-based AML solutions look like? How do you see the technology evolving in the future?
I think machine-learning will be increasingly adopted as the benefits become more well known as opposed to the ‘terminator complex’ that tends to often to spring to mind when people hear about AI.
The space we are in right now is complex. We are constantly stopping people from doing bad things. It is important to remember that these guys bring a lot of resources and power into play. They are also constantly trying to dodge our efforts.
So, it only makes us believe that we should keep improving and getting better at using advanced technologies, such as artificial intelligence. I think the future in this space is tremendous. In the next five years, we are going to see a big shakeup in terms of who are the new market leaders. More and more players are coming up — offering advanced solutions that can be used by financial institutions without sacrificing their gains. In my view, this is the way to go. Mostly, we look forward to supporting our clients in redefining what it best looks like when it comes to keeping criminals away from the global financial markets.