Research-2

Idenfo – Artificial Intelligence: An Effective New Tool In Fighting Financial Crime

a look at the impact of AI on AML and KYC processes

by Zorez Haider

Legal and Compliance Analyst

Artificial Intelligence (AI) is fundamentally changing the way we are fighting financial crime. New technologies are allowing institutions to gain a better understanding of their customer’s activities, helping to more quickly and efficiently fight financial crime. Idenfo is uniquely placed to take advantage of such technologies by providing its clients with a tech-driven solution to their Anti-Money Laundering (AML) and Know Your Customer (KYC) needs. We use AI automation to provide our clients with document verification, name screening and risk rating capabilities which can be tailored to their unique requirements, helping them to fight financial crime and steer clear of financial penalties from regulators.

AI and Document Verification

Document Identification (ID) fraud is an increasing problem in today’s highly digital society. Through the use of AI software, it is now possible to quickly and seamlessly match an individual’s document ID photograph with their likeness, verifying their identity. Historically, this would have to be done manually in a time and labour intensive process. However criminals are developing new, ever more sophisticated methods to get past AI-based document verification. At Idenfo, we use a ‘liveness check’ which involves a short video recording of the customer. This allows us to match the customer’s ID photograph with a much higher degree of accuracy.

AI can also be trained to detect signs of forgery or tampering to determine the authenticity of a document. Additionally, it is now possible to analyse a document’s expiry date and document number against the standard form of government-issued documents in a jurisdiction. AI can also be trained to recognise holograms and micro prints on most government-issued documents, helping to more accurately verify their authenticity. These are some of a number of factors where AI can be used to quickly and accurately ensure the legitimacy of a document. AI-driven document ID verification can reduce business risks by scanning a high volume of documents with a high degree of accuracy as compared to the traditionally manual, labour-intensive document ID verification process.

Name screening

Name screening is an important part of running KYC and AML checks. One of the key difficulties is a high level of ‘false positives’, due to many people sharing names with individuals on sanctions lists, such as the EU Terrorist List. Additionally, individuals on sanctions list are more likely to submit a variation on their name in order to pass through name screening checks. AI can help streamline the tenuous name screening process by scanning sanctions list more quickly. Similarly, AI can be trained to flag potential name variations by using a technique known as ‘fuzzy matching’ to capture any potential name variations used by sanctioned individuals to slip through AML checks. Additionally, name screening can be used in conjunction with other checks to reduce false positives where a customer shares a name with sanctioned individual. This is discussed in more detail below. AI-driven name screening solutions can be tailored to match an institution’s risk appetite, helping to reduce ‘false positives’, and helping companies to avoid damage to their reputation from unintended sanctions violations.

AI and ‘risk ratings’

AI can be used to deliver holistic AML and KYC solutions more efficiently and quickly to clients. By using document ID and name screening checks, in addition to the analysis data points such as an individual’s country of residence, nationality, and career, AI-based solutions can help deliver a simplified ‘Risk Rating’ for a potential customer. This risk rating can help an institution determine whether on-boarding a customer opens them up to a potential sanctions violation in a quick, simple and accessible way. This risk rating, a key part of Idenfo’s product, can be customised to suit an organisation’s requirements. For example, an institution in Pakistan with a primarily Pakistani customer base may have over-inflated risk rating due to the Pakistan’s status as a high-risk country. This can be adjusted to a level suitable to an institution, and a higher threshold can be applied for labelling an individual as ‘high risk’. Machine learning algorithms and natural language processing techniques can then be used to further refine the AI software’s ability to capture and understand data points relevant to providing an individual with a risk rating. This is an effective way to grow an institutions AML and KYC capabilities while staying within parameters relevant to their business; a truly individualised and customisable solution.

Conclusion

The most important factor to consider is that AI can be moulded to match a company’s unique AML and KYC requirements, allowing a flexible solution to tackling financial crime. With penalties and regulations growing as the financial industry enters the digital age, it is important that institutions embrace technology to meet the increasing regulatory obligations they face. AI techniques are heralding a radical change in how AML and KYC checks take place, helping to provide cheaper, more efficient and accurate solutions which can be shaped to meet an organisation’s individual needs.