Fighting financial crime by revolutionizing AML compliance: our investment in Sentinels
We are hugely excited to announce our investment in Sentinels, a Dutch AI-powered transaction monitoring start-up tackling the issue of money laundering and other illegal financial activity. As we proudly lead this seed round of €5.7m, we are joined by a group of high-profile angel investors and industry leaders: among others, we are pleased to partner up with several members of the early Adyen team — including Thijn Lamers (former EVP Sales), Maikel Lobbezoo (former Head of Growth), who also participated in this round.
The gap between what financial criminals are capable of and what compliance teams can do to stop them has been widening…
Globally, around USD 1–2 trillion is being laundered per annum, representing c. 2–5% of global GDP. What’s more, only 1–3% of laundered funds are identified and possibly stopped. With such dazzling numbers, the scope of the problem necessitates strong supervision and it is hence not surprising to see headlines being dictated by massive fines imposed on financial institutions. From Commerzbank to Goldman Sachs and from Deutsche Bank to ING, all these seemingly sophisticated banks have fallen victim to massive AML compliance scandals over the recent years.
The irony, however, is that the global spend on financial compliance was estimated at a staggering USD 180.9bn in 2020[1]. So this is certainly not a matter of under-investment into tackling the problem, right? However, it all becomes clear when you double click on what the billions of compliance spend accrues to; the spend is dominated by throwing more people and legacy technologies (if tech at all) at the problem.
Now, banks do have the capacity and financial muscle to throw a massive amount of people at the problem. Fintechs (such as PSPs, neobanks, neobrokers, remittance PSPs, money transfer operators, crypto-exchanges, e-money institutions etc.), however, do not have these budgets, and it’s hard for nascent companies to defend themselves. Although fintech has disrupted almost every corner of the financial industry, compliance innovation hasn’t had the same impact. Criminals are leveraging new capabilities and are keen to try laundering through fintechs. Supervisory pressure is following suit. Who will defend them?
Rule-based approaches require augmentation with ML
Transaction monitoring tools for AML compliance have been a mainstay process within financial firms for many years, using a rule-based approach to identify certain account-based activities for investigation and disclosure of suspicious activity. However, such an approach is dichotomous, static and cannot cater to today’s dynamic world of transactions and payments. Following increased regulatory pressure, compliance teams have reacted by creating more rules, to (show the regulator they are working to) identify more potentially suspicious activities. The result? An over-burdened with false alerts. Consequently, actual cases of AML go unnoticed (false negatives not getting due attention), culminating into gigantic fines for institutions and even personal liability for executives. Financial institutions cannot bear the burden of linearly increasing compliance costs and threat of fines for much longer, and fintechs, looking for profitability certainly can neither .
The Sentinels’ game changer: ML of the pragmatic sort
Sentinels has built a solution that augments business rules with Machine Learning (‘’ML’’) algorithms to detect suspicious transactions with a higher level of accuracy. This combination of business rules and (unsupervised) machine learning can significantly reduce the number of false-positives from 96% to up to 75%, a game changer in the AML compliance industry. Sentinels is able to achieve greater accuracy, by starting with richer context. The solution unlocks internal data siloes and augments it with external sources, all via simple APIs. The Sentinels ex-post transaction monitoring solution runs a holistic analysis to understand the nature of transactions processed by a client (i.e. a merchant in the context of a PSP) and how their behavior compares against the expected transaction profiles of the client in question given market trends, seasonality, peer group averages, etc. so as to surface suspicious activity. The platform also allows fintechs and banks to evaluate networks of relationships of transacting entities within the broader client network through a graphical network analysis interface. Moreover, while the platform is already extremely powerful as it stands now, Sentinels will provide even more utility for its customers going forward, as it moves into real-time transaction filtering and builds upon its mission to provide continuous client surveillance.
Besides empowering compliance teams with an unprecedented AI-powered analytical platform, Sentinels also goes a couple of steps further by simplifying existing workflows within compliance teams. It provides a step change in efficiency in case investigations and increased control with reporting tools. Sentinels’ solution is compliant with the highest level prescribed by regulatory bodies, providing a ‘compliance in a box’ solution to companies which otherwise need to abide by thousands of pages of regulatory information to stay compliant.
Investment Thesis
At INKEF we couldn’t be more excited to back Joost van Houten and his team. We would like to cite the following reasons for investing:
Immensely talented team with deep roots in AI and ML
Sentinels was created within Slimmer AI, a venture studio focusing on applied AI ventures in close association with the University of Groningen, a top-rated AI faculty in Europe. The extremely talented team of engineers is complemented by a strong team of entrepreneurs under the leadership of Founder & CEO Joost Van Houten. In our interactions with Joost, what we found was an enormous level of ambition, deep domain knowledge and a natural charisma that has already helped the company attract strong leaders from the fintech & payments industry such as Ibrahim (CTO) and Sumit (CCO). Last, as part of this seed financing round, Sentinels has surrounded itself by an incredibly strong network of industry leaders.
A pragmatic approach to AI
An important reason why the compliance industry is riddled with legacy technology is that the typical ‘black box’ ML models are not suitable for such a highly regulated industry. Sentinels’ has built a three layered technology for transaction monitoring with business rules, unsupervised learning, and network analysis that provides a transparent and continuously improving detection approach. The focus of Sentinels has been on augmenting the existing procedures in place with the breakthroughs in technology, while staying compliant with rapidly evolving regulations.
Large market with strong tailwinds
The market is growing rapidly as ecommerce and online transactions proliferate, more fintech companies emerge, regulatory pressure mounts and more fines are to be expected. The company is already tackling a multi-billion dollar problem and its serviceable addressable market will only expand as it moves into adjacent fields (no, we won’t be divulging too much at this stage 😊)
Early traction and trust from leading fintech’s and financial institutions
Sentinels launched its MVP in October 2019 and has already signed several high profile customers ranging from small all the way to large fintechs (such as unicorn Mollie, which was a launching customer), as well as more traditional banks. The trust signaled by these and many other companies with whom discussions are under way is a testimony to the level of maturity of the team and the product they have built in a very short period. And what we already knew about this industry is that if you can earn the trust of compliance officers, with regulators breathing down their neck, the rewards are sure to follow as this is one of the most tightly-knit industries that we are aware of.
We are beyond thrilled to join the Sentinels team on this exciting journey!
[1] LexisNexis® Risk Solutions, True Cost of Financial Crime Compliance Global Report
