This finds a constant mention in our customer conversations, with the stakeholders repeatedly pointing to the challenges that multiple global regulatory norms pose. There is a tremendous amount of back and forth between local and centralized KYC teams on validation of KYC reports.
Spend on AML and compliance
Our discussions with industry experts suggest that global spend on KYC is currently well above USD 6 billon and growing every year. While the full-time staff strength for compliance varies from bank to bank, the numbers have been increasing in response to complex regulatory requirements and fines imposed on the financial services industry. JP Morgan, for instance, added 4,000 additional compliance staff in 2013, HSBC took on 3,000 additional people after it was fined USD 1.9 Billion for money laundering, taking its total compliance staff strength to 7,000 in 2015. To be sure, these are not isolated cases; others including Deutsche Bank, Credit Suisse, UBS and Citigroup have stated significantly increased regulatory and compliance spending in last few years. Our research suggests that as of now, technology is about 20% of overall AML compliance cost.
To top this, financial institutions spend significant amounts on subscriptions to data providers for verifying customer verification. This is despite a lot of this information being available publicly, such as in the annual reports on the company website, SEC filings and on third-party sites such as Wall Street Journal and Stock Exchanges. Analysts must painfully go through these sources to obtain the pertinent data points.
Technology enabling KYC automation for enhanced compliance
While there are industry specific solutions in the regulatory compliance space, accelerating risk mitigation with productivity gains while enhancing customer experience remains an imposing objective.
The present offerings related to AML and KYC space are categorized into one of the below:
- Software that provides the API and algorithm to capture key attributes
- A database that is used to verify the information from a selected source
This is a crowded landscape with most players’ capability being around centralizing collection and management of KYC data or a utility that allows firms to update and/or access information.
We regard these approaches as useful for temporary boosts, but inadequate in the bigger scheme of things. The current manual operations need rethinking; at these volumes, manually checking each client’s credentials and recording the data cannot work. Accommodating large-scale remediation while maintaining accuracy and timelines requires smart automation, with manual intervention only on a case-to-case basis.
Capabilities such as machine learning, contextualized search, intelligently mimicking human actions and handling unstructured data can help scale the operations effortlessly.
Application of these capabilities in the KYC world includes identifying relevant documents, understanding, and extracting attributes from these documents and archiving information as evidence. Additionally, contextual understanding implies negative news about an individual/entity can quickly be brought to the forefront and discrepancies automatically red flagged for investigation.
Taking away the mundane and low-value adding activities of data search, aggregation and extraction multiplies the team’s productivity and improves focus on critical reviews.
Another challenge that FIs face is handling and identifying unfavorable information about a person or entity. For this to be effective, compliance professionals need to be able to filter through this overwhelming amount of data from various news sources. This is easier said than done—and automated negative news discounting utility can help them focus on areas of higher risk. Advanced technologies do provide a solution, by discarding the irrelevant news, while ranking the top few negative news articles for consumption of investigators.
An essential cost for the FIs, of course, is access to accurate data. They rely on data providers for information such as company’s ownership structure, financial data, controller information, address, etc. AI steps in and does this efficiently through credible public resources, drastically cutting down dependence on paid data sources and subscriptions both for Enterprise KYC (E-KYC) and Retail KYC.
Our vision: New era of KYC
We envision an enhanced offering that holistically covers peripheral business areas of KYC while expanding the geographical countries.
Risk-based transaction monitoring to flag money laundering alerts with the use of machine learning. The current mechanisms flag a large volume of transactions as suspicious, with false positive rates hovering around 90%. Reducing false positives can be a big boost to bandwidth strapped investigation teams, significantly improving their performance by probing alerts that matter.
In addition to adverse news search, checks related to politically exposed persons (PEPs) and their relatives, and sanctions list screening of beneficial owners are conducted as part of the KYC process. Comprehensive screening can decipher complex structures that typically carry the highest hidden risk. This can be a labor-intensive exercise needing continuous monitoring, and manual processing can open the firm to potential risks.
Also, we see the need to bring additional data sources aboard to enrich the quality of checks. This includes bank’s specific sources, or sources specific to retail vs. institutional KYC, or are relevant for a specific country. These will make the checks more diversified, precise and add a layer of validation.
In the next paper, we will find out about the impact of intelligent automation on bank’s key metrics, on KYC analyst’s productivity and the challenges we faced in the automation journey.