Silent Eight’s Markiewicz: "Costs force fighting fraud to shift from people to AI"

• While banks have tools in place to fight financial crime, many are manual and expensive
• Using AI-based tools can help banks escape the spiral of escalating costs
• While returns may only become visible during high-profile events such as investigations, banks are also calculating ROI and finding positive payback

                Banks are doing what needs to be done to prevent money laundering and fight financial crime, said Silent Eight CEO Martin Markiewicz. “The way they do this stuff is modern. Systems are built around rules-based engines. You put in manual work flows, you use people, you try to sort it out.”

                What many banks have been doing to deal with fraud, Markiewicz said, is to use rules-based systems, many of which were developed “a long time ago”. The challenge, Markiewicz said, is that these supposedly-modern systems “were designed a long time ago, before AI started being useful. There is a lot of manual effort around fighting financial crime.” Even though the systems may seem modern, they require an ever-expanding pool of people to run them and they’re not fully dealing with the even-more-advanced tools criminal that use. The result, he explained, is that “fighting financial crime costs more and more every year. The manual solutions are not scalable. If you have a thousand people and the world changes, doubling the scheme and hiring an extra thousand people is not easy to do. The process comes with higher costs. They are flagging more, preventing bad situations, at a cost. The better they get, the more everything costs.”

                A preferable alternative to escape from this spiral, Markiewicz believes, is using cutting-edge technology. “For the banks to keep up, you have to arm yourself with better weapons. You want to do this in a way that you are still profitable. There’s nothing gained when you don’t address all these issues.”

                What is far better, he explained, is using artificial intelligence (AI) “to create an improvement to the process that already exists.”. “If you look around, not many banks are doing smart things, using the newest available tools. Not everyone around them has it. It’s early stage. You have to design this technology, use it to build a good solution, remembering what we can do, cannot do.”

                In its work with Standard Chartered Bank, for example, Silent Eight has developed screening optimisation to apply machine learning and natural language processing techniques to improve name screening, whereby customer and related party names are matched against watchlists. By analysing historical decisions, the optimisation engine learns to replicate the assessment in the same way that a human analyst would respond.

                The improvements are significant, helping to reduce costs, increase accuracy and make the process more effective. “No one wants to go through all this just to gain 3%,” Markiewicz observed. “It has to be significant.” Beyond the financials, banks are using AI to keep the financial system safe and keep themselves out of trouble.

                Although banks are working to measure results, analytics can be challenging. One obvious metric is return on investment (ROI). While they can measure that cost-benefit, Markiewicz said the hard part is around measuring AML results. “They may not see the return until they spot it during an investigation, or someone gets arrested, or someone says the bank is financing North Korea.”

                “If you blindly introduce a solution, it will work great on paper,” Markiewicz opined. However, “it won’t work in real life.” Banks need to test and measure impact of AI on top of what they’re doing, and to satisfy all the stakeholders. “The biggest banks across the world, stakeholders are huge, you have to make sure everyone is okay.”

                  “Everyone knows they need to move forward against bad actors,” Markiewicz said. And “if you look at the cost of compliance, what banks are doing in a traditional way, it looks like spending more and more money.”

                While it is still early days, banks that are starting to use AI solutions are making changes they need, more cost-effectively, and starting to measure the results.