Fighting against financial criminals requires extensive teamwork, intellectual machine learning and a new concept pivotal to AML firms, data exchanging within the private sector – this is a firm belief of Salv Co-Founder and CEO Taavi Tamkivi.

Tamkivi discussed with Payment Expert how he helped shape the AML and anti-fraud teams at Skype and TransferWise. Tamkivi spoke with Payment Expert about the fundamental factors for companies to fight fraud, the importance of data sharing and the evolving nature of financial criminals. 

Payment Expert: What are Salv doing differently when it comes to KYC and Risk Scoring checks? 

Taavi Tamkivi: Under the AML legislation, KYC it’s a separate process. Salv is focusing on several processes that fall under the legislation umbrella – risk assessment, sanction screening and data exchange between the banks and fintechs. 

The main problem that companies of any size are experiencing right now is the growing weight of the regulatory responsibility and the lack of efficient technologies to help them stay compliant. As a result, large institutions end up hiring hundreds or thousands of people to work manually, often checking the false-positive cases, flagged by the inefficient tools that the institutions are currently using. 

This doesn’t help fight financial crime and the countless hours spent on cases, where the actual crime rate is incredibly slow, could be better spent elsewhere. Salv is changing this by helping institutions to define rules and algorithms so that they can correctly identify more possible crimes. We improve the utilisation of data, using it more flexibly, building adjustable rules on top of this data and making it user-friendly. 

At the moment, the banks, fintechs and other players in the space often don’t have efficient processes in place to inform law enforcement whether a suspicious user or a transaction ended up being a ‘dirty player’ or an innocent one. Salv aims to resolve this with data exchange and collaborative investigations between the institutions. As each party will have data on the suspicious transactions and Salv will provide them with an intelligence-sharing platform AML Bridge, that’ll enable them to directly share insights or get answers to questions, allowing them to complete the data. 

PE: From your experience, what are some of the vital functions of AML systems such as yours for startups or financial institutions that are yet to implement such systems? 

TT: Companies need to govern a set of procedures that are defined by the laws, regardless of whether they are large banks or small fintechs, everyone needs to follow the same legislation. All these companies want to scale, and if they have a large volume of customers and transactions they need to use some technical systems, if the volumes are more humble it can be done manually. As our customers process lots of data, they need to have a system in place which is defined by the laws. 

But it’s not enough today to have only great screening, KYC or risk scoring tools in place. To better tackle AML institutions need technology that enables them to share intelligence and collaborate. There are already countries that are innovating in this field, with Singapore, Netherlands and Estonia being great examples for connecting banks to share intelligence or pool data for better analysis.

Salv’s contribution to this collaborative crime-fighting is our secure inter-institutional double encrypted communications platform called ‘AML Bridge’ that enables the immediate exchange of information between two parties.

PE: Were there any big takeaways or learning experiences you took from building fraud detection teams at Skype and TransferWise which you then applied to Salv? 

TT: One crucial thing: when people are building up the organisation to stop fraud, they often keep the data analytics team separate from operations, which is again separated from technology. In my case at Skype and TransferWise, people in operations were sitting next to the data scientists, who had the big picture on data points, and engineers and product teams were sitting next to each other. This synergy worked brilliantly in both of these companies. 

Unfortunately, many big companies do not do this; their operational team is very far from their data analysts and technologies, and that’s what we keep in mind when we build our products for our customers. We have a robust compliance expert team helping our clients build rules to advise them and discuss missing opportunities. 

The science team is helping our clients to analyse their data on a large scale. Finally, we have product and technology teams, so we are trying to serve a combination of these three critical sources to our customers, but also without the customers, this is how an ideal crime-fighting team can be built. 

PE: As fraudsters and financial criminals develop new ways of fraud and money laundering, how do Salv acclimate to these new threats? 

TT: Criminals are still working at large networks, sharing data like stolen documents or bank account details there using different institutions and methods like fiat currency to cryptocurrency or other assets. They control the network from their home like supply and demand, selling bank service information to one another. 

The law-abiding institutions are not working in these kinds of networks, predominantly operating in silos. They only have the data points in their own machines which they can see and analyse, but it is very similar to puzzle pieces – trying to guess what is the big picture. Criminals have this full picture but crime fighters do not. In order to help them, we also need to know what is happening around them, so we are building a network and intelligence-sharing platform for crime fighters. It takes a network to beat a network. 

The three current important influencers in the markets are the evolution of financial technologies, the impact of COVID, and the Ukraine-Russian war. 

PE: Public-private sharing seems to be a key factor when it comes to companies monitoring and detecting financial crime. What more do you believe can be done to make it even more seamless for companies to share information with one another? 

TT: It is a very immature process right at the moment. Every financial institution that reports to the FIU (Financial Intelligence Unit) reports cases believing ‘this is a suspicious transaction or behaviour’ and hands it over to the police. But actually, they never know whether it was or wasn’t suspicious or even  necessary to report, thereby unintentionally creating extra work for the police force. Problematically – there is no feedback loop. 

This is why I believe in the private sector, and the power of data sharing between financial institutions. Last year we launched the AML Bridge in Estonia with a few participating local banks, supported by the FIU and FSA (Financial Services Authority) to test our collaborative crime-fighting platform AML Bridge.  

Initially, the participating banks were quite suspicious, but after seeing its value they became quite excited – as more and more messages were sent it helped build up new use cases of fraud and AML. The rest of the banking market heard of this new intelligence-sharing network, and more wanted to join in, so it took six months to cover the whole banking market [in Estonia], and banks are usually relatively slow to move. 

This kind of growth shows how big of a problem there is, and once they’re shown the way how willing the banks are to start using it. Now we are kicking off the same ideas in other countries in Europe, and we hope to see the same growth once banks and fintechs realise that it is bringing tangible benefits. 

PE: Are there any new innovative approaches/methods Salv or you have observed that could be the future of taking down financial crime a lot quicker and efficiently?

TT: I think there are three trends in the market; we are betting on one of them but not all. One is the use of machine learning and AI technologies. Large institutions have vast amounts of data which can be too complex for human beings to understand, so machine learning can help find suspicious patterns that humans would not be able to. But this is not the core problem of the AML space. It is also important to receive feedback from machine learning, and as I’ve mentioned before – this is not always the case. 

A second trend is the visualisation of data points that people want to see huge complex graphs of the users’ data. This is useful for investigations when the officials want to map out complex data points, accounts and assets. 

And the third one, the one I believe the most in, is making data from other institutions available for the crime-fighting teams. I am happy to say that the industry body The Financial Action Task Force (FATF) is betting on this. The next president was calling out the big techs, banks and more, advising more private-to-private data exchange to make the crime fighters smarter.