Canadian FinTech startup Koho is leading the way in leveraging generative artificial intelligence (AI) to transform its anti-money laundering (AML) and counter-terrorism financing efforts. The company has developed an internal tool that significantly accelerates the investigation process, reducing the time analysts need to complete a typical investigation from 95 minutes to just 35. By automating much of the repetitive work associated with these investigations, Koho’s AI tool allows investigators to focus more on critical analysis and decision-making. Despite the automation, the company has made it clear that the tool is not designed to replace human oversight but to serve as a supportive assistant to streamline workflows and enhance productivity.
David Kormushoff, Vice President of Technology and AI at Koho, discussed the tool’s impact during a panel at the Amazon Web Services (AWS) Re:Invent conference, highlighting the potential for even further reduction in investigation time. He shared that Koho aims to bring investigation time down to just five minutes, with human investigators continuing to check the tool’s work for accuracy. This goal reflects the growing potential of AI to automate routine tasks while keeping human experts in the loop for critical oversight, ensuring that AI systems do not act in isolation but instead augment human intelligence.
Koho is regulated by the Financial Transactions and Reports Analysis Centre of Canada (FINTRAC), which oversees the detection and reporting of suspicious financial activities. This includes identifying and reporting instances of money laundering and terrorism financing. However, the challenge for Koho and similar financial institutions is the sophistication of bad actors and the sheer volume of transactions processed every day. Financial criminals are increasingly adept at layering transactions across various platforms and institutions, making it difficult to detect suspicious behavior through traditional means. This results in a backlog of investigations in many financial institutions, as risk and fraud departments struggle to keep pace with the volume of data they need to analyze.
As Kormushoff noted, Koho’s target demographic—individuals living paycheck to paycheck—is particularly vulnerable to these criminal activities. These individuals are often the target of scams, such as fraudsters calling elderly people and convincing them to purchase gift cards as a form of payment, which is not only a form of fraud but also money laundering. In these cases, the amount of time required to investigate each alert manually is a major hurdle, as investigators must go through years of transaction data, cross-checking against public sources, and compiling their findings into a report for FINTRAC.
Koho’s generative AI tool now automates much of this investigative process. The AI is capable of reviewing transaction histories, flagging suspicious patterns, and searching public databases for additional information about the individual involved. It then generates a report that cites all the sources used in the investigation, providing transparency and allowing the human investigator to validate the findings. While the tool is incredibly efficient at reducing the time and effort needed to complete investigations, it does not replace the analyst’s role in reviewing and confirming the AI’s work. Instead, it acts as a valuable assistant that speeds up the process without compromising the quality or accuracy of the investigation.
One key aspect that sets Koho’s AI tool apart is its focus on privacy and data security. Kormushoff emphasized that the tool is not trained on customer data and only receives information relevant to the specific investigation at hand. This approach ensures that no sensitive customer data is used outside the scope of the investigation, protecting privacy while still allowing the tool to perform its function effectively. The AI system is designed to maintain compliance with both regulatory standards and privacy laws, ensuring that Koho remains fully transparent and accountable in its use of AI technology.
The AI tool’s ability to automate much of the investigative work not only improves efficiency but also enhances the overall security of Koho’s operations. By reducing the time needed to identify suspicious activity, the tool allows Koho to respond more quickly to potential threats, reducing the window of opportunity for criminals to exploit vulnerabilities in the system. This makes it easier for the company to protect both its customers and its financial ecosystem.
Furthermore, Koho’s innovative use of generative AI underscores the company’s commitment to harnessing cutting-edge technology to address some of the most pressing challenges in financial crime prevention. The tool not only helps Koho comply with regulatory requirements but also enhances its ability to protect its customers, particularly vulnerable populations, from fraud and financial exploitation.
In the long run, Koho’s use of AI in the fight against money laundering and terrorism financing could serve as a model for other financial institutions, illustrating how technology can be used to improve operational efficiency and reduce the risks associated with financial crime. As AI tools continue to evolve, it is likely that their role in financial security will only grow, offering even more sophisticated ways to detect and respond to illicit activities while maintaining privacy, security, and compliance with regulations.