Notion Capital helps Resistant AI build the next generation of AFC
We are proud to announce that Kamil from Notion Capital joins our existing investors—Tom & Vidu from GV, Jan & Erin from Index, Ondrej and Vlado from Credo, and Reshma and Carlos from Seedcamp. Feel free to consult our press release for the technical details on the funding, but we would like to explain why we chose to work with the Notion team.
Not only do we share a common culture and legacy in network and computer security with the Notion team, but as the founders of Message Labs, they paved the way before us in launching a completely new product category into an emerging, rapidly evolving B2B market—the content security market—and successfully leading the company towards a major acquisition by Symantec in 2008.
The new context of financial crime
Message Labs was founded to address an emerging tangible threat: the propagation of malicious code hidden in internet-delivered content. Resistant AI was started in similar circumstances: the rise of fraud and fincrime facilitated by digital technology. The wave of APP fraud, ransomware money laundering scandals, wash trading and market manipulation in crypto, and the epidemics of identity theft, synthetic identities, and other attacks related to application fraud are all profoundly transforming the financial industry.
These crimes are just symptoms. We are living through a convergence of financial, fraud and cybercrime: it turns out that when you digitalize a bank and have algorithms take decisions instead of people, the algorithms become targets of fraud instead of people. And so the techniques originally developed in the cybercrime domain are being exploited to directly attack these new code-based, automated financial processes.
Time is another factor impacted by digitalization: banking operations now take seconds instead of days, and so does the crime. Fully automated crimes can now be completed in seconds—far less time than it takes to even notice an alert in a traditional financial crime prevention system. To make the problem even more difficult, the complexity and sophistication of these crimes goes far beyond simple payment fraud: many of them can only be identified as criminal when a broader context from multiple distinct banking systems—KYC, transaction monitoring, risk scoring, and others—is taken into account.
Yet another level of complexity comes with scale. Traditionally, financial crime depended on black markets for not-quite-scalable (human) services such as money muling. Digital transformation changes the fundamental economics of financial crime. Robotic, fully digital money mules interacting with networks of fake businesses make financial crime far more scalable: smaller amounts can be used for money laundering, spread over a higher number of fully digital identities that would only be criminally used a few times, obscuring links and making the traditional detection methods ineffective.
Finally, AI is coming in to act as an accelerant in all those trends, making everything—both services and crime—faster, more scalable, and more accessible. But it's also providing the tools needed to fight back. It is at this intersection of fraud, fincrime, and cybersecurity; of AI, automation, and traditional finance; of the physical and the digital, that a new field of anti-financial crime is emerging—one that Resistant AI was built to serve.
Customers of financial service providers don’t care how we protect them from novel threats or how we categorize these threats. They want their data and money to be safe. And they are increasingly backed by lawmakers and regulators.
How Resistant AI changes the game in financial crime fighting
What makes the Resistant approach different is inspired by our background in cybersecurity. We do not believe in single answers to large, complex problems—instead, we layer best-in-class methods together for decisive robustness. Nor do we believe in building AI “solutions” that drown customers in complexity—only to create exploitable opportunities for criminals. We build narrowly focused AI products that offer reliable performance, understandable outcomes, and repeatable deployment and value delivery.
Those product layers start from the beginning of the customer journey, at the intersection of the physical and the digital, and continue on into the ongoing and evolving relationship of that customer, at the intersection of the digital and the transactional.
Protecting onboardings
At onboarding, we prevent the creation of robotic accounts based on stolen or synthetic identities of companies and individuals by analyzing everything from the documents they submit to the behaviors they exhibit, and the information they provide in forms. We surface and highlight anomalies, whether they are deviations from the norm in the individual, or statistically impossible repetitions across accounts. If the core of any scalable criminal endeavor is repeatability, our multifaceted detection makes it impossible for criminals to be financially viable.
Protecting relationships
AML transaction monitoring is a chore, often disconnected from the rest of the customer risk assessment activities and underpinned by an outdated rules-based approach that generates more noise than findings. However, changing those systems is equally painful. We provide an overlay system that makes these processes far more efficient by intelligently prioritizing all of the alerts from the underlying system, adding much-needed context, and highlighting the critical ones that truly need the attention of highly trained financial crime investigators. We also use our access to documents, behaviors, and transaction data to find new threats undetected by the underlying system.
Protecting identities
To accomplish the above goals, we are building and maintaining a customer-centric model of all documents, behavioral characteristics, digital traces, and financial transactions conducted by the customer on the financial platform. This model is then used to make fully contextualized, highly precise decisions in the various domains of compliance, risk, and operational decisions, minimizing routine work and concentrating attention on the high-risk cases, provided with all the relevant details.
Resistant AI is on a mission to protect the financial technologies powering the embedded finance ecosystem from the new generation of threats. By providing adaptive, AI-powered security that protects financial systems, we enable our customers to build increasingly sophisticated and embedded financial services accessible to everyone—except fraudsters and the financial criminals.