Netflix is the world’s largest streaming provider and has become particularly popular in recent times. It gained about 10% new subscribers in the first quarter of this year – twice as many as forecast. It’s certainly convenient. The last season of your favourite series is barely over before a new show is suggested.
How does Netflix do this? Machine learning holds the key
Machine learning is an application of artificial intelligence that allows systems to generate knowledge from experience. The machine learning algorithm learns from historical data and is able to make recommendations and predictions based on this information. This means that Netflix can use our viewing history, and the viewing history of other users, to suggest other films and series that we might enjoy. The system is continuously learning and gaining intelligence. The algorithm behind the system is trained using a wide range of techniques, including natural language processing, which can be used to process and understand text.
But what does that have to do with compliance?
From Facebook and Instagram to Netflix and Amazon – artificial intelligence is ubiquitous in our daily lives. And it is gaining in importance in the financial industry, too. In principle, machine learning can be applied to any area in which large quantities of data are processed. In compliance, and in the administration and management of KYC data in particular, the possibilities offered by machine learning algorithms are very promising.
KYC data is a stumbling block
High-quality KYC data is not only expected by international regulators, it is also essential for processes such as client risk classification, transaction monitoring, media screening and periodic reviews. A series of money laundering scandals have also shown that client information and the monitoring of relationships and transactions are often inadequate, despite the efforts made. The resulting financial damage and reputational risk have become a reality for many financial services providers. This pressure is increasing as the requirements for compliance with anti-money laundering provisions become ever-more complex, requiring continuous efforts from front-office and compliance teams.
A new era?
The use of machine learning algorithms offers an opportunity to assist with the qualitative assessment of client information, and to identify inconsistencies and suspicious behaviour. For example, a machine learning algorithm can flag transactions that deviate from the client profile and behaviour, and show whether a name screening match actually relates to the client concerned. Genuine risks can thus be identified more efficiently and compliance resources can be used in a focused manner. In addition to offering efficiency gains, machine learning algorithms pave the way to increased compliance security and more robust client data quality.