Financial institutions are increasingly recognizing the importance of KYC (Know Your Customer) as a crucial process to protect against financial fraud, terrorist financing, and money laundering. KYC involves gathering essential customer identity and address information. Regulatory bodies have been imposing strict penalties on non-compliant organizations, prompting many to invest in advanced financial transaction surveillance systems. However, the industry faces significant challenges in terms of the vast amount of data, its rapid generation, and the complexities arising from multiple and non-standard formats.
What is KYC?
#KYC is a vital process conducted by relationship managers to ensure compliance and prevent fraudulent activities. It is the customers due diligence ( #CDD ) and bank regulation that financial institutions and other regulated companies must perform to identify their clients and ascertain relevant information pertinent to doing financial business with them. It involves conducting background checks and verifying customer identities.
When a financial institute works on how to know your customers, it is crucial to make sure that it is establishing a trustworthiness of the banking relationship and effectively managing the client's assets, including executing transactions involving financial products.
Why do we need to do KYC?
Know Your Customer (KYC) data can help banks reduce risk by providing a means by which to identify customers who are likely to default on a loan. Most banks around the world have implemented account monitoring mechanisms to guard against financial fraud, money laundering, and terrorist financing.
Let’s look into more data.
The International Monetary Fund (IMF) states that laundered money accounts for 2%–5% of the global GDP. Regarding to a market survey, reaching data from 210 financial institutions in six Asian countries (China, Hong Kong, Indonesia, Malaysia, Singapore and Thailand), research estimated that the total annual cost of complying with the law at around US$25.3 billion in 2018.
When we look into China, the Chinese market has a large and complex financial system including some of the world’s largest commercial banks. Credit card fraud is the most common fraud type that commercial banks in China face. In every 10,000 credit cards issued by the Postal Savings Bank of China, there were 0.83 cards related to credit card fraud, ranking first among the leading ten credit card issuers.
Who is in response to gatekeep the KYC and compliance process?
Today, FATF oversees and is a flag bearer for money laundering, terrorist financing, and combating arms proliferation through the creation of regulatory and process-driven measures for the international financial system.
FATF in its current structure reviews the robustness of every country’s ability to combat money laundering, counterterrorism efforts and make recommendations to ensure that their financial system and regulatory environment are more efficient. It primarily assesses the member countries on two broad parameters:
Effectiveness—ability to have a robust framework that protects the financial system from financial crimes. It has 11 broad areas on which FATF assesses the effectiveness of the financial crime framework.
Technical compliance—this refers to a list of 40 compliance items that needs to be met. It contains assessment on law, regulation, and legal instruments that can control terrorism or its proliferation and money laundering.
How does KYC work?
Traditionally, this involved extensive research, including examining corporate ethics, financial investments, press releases, annual reports, and even leaders' social media presence. Apart from name matching, the key idea of KYC is to monitor transactions of a customer against their recorded profile history on their accounts with peers.
In the digital transformation era, many of the FinTech companies, like iFinGate, is now providing AI solutions to automate the KYC process. Not only could it enhance the efficiency but also the accuracy of the process when doing Know-your-customers.
What are the problems on doing KYC in present-day practice?
As a professional in the banking industry for decades, we know that you may hesitate in changing the old practice to a new norm. But have you ever been told about the potential crisis that your institute might be coming across due to the old practice?
According to scholars (Sai, B. D., Nikhil, R., Prasad, S., & Naik, N. S. ,2023), there are significant loopholes in the present-day KYC model. In the present-day flow for document verification used for KYC verification, many involved parties hold your documents, which has a high risk of misusing them. There are high chances of any party with compromised ethics entering into this KYC network, getting your documents and misusing them. In other cases, there is also a high chance that any party with compromised security entering into this network of KYC can get their databases hacked, which can lead to a data breach.
Even considering a party with high ethics and highly secured databases, it costs those parties a high charge for the KYC verification and maintaining the security of the databases. These charges can burden companies that verify their customers with the KYC verification process. It’s not only the companies verifying the documents that are meant for this financial burden. Even the government bodies are victims of this financial burden. The government even should spend to secure their databases.
Why should we apply AI solution in KYC?
Applying AI technology to KYC (Know Your Customer) processes offers several advantages, particularly in relation to the 3-Vs: Volume, Velocity, and Variety of data.
Volume: KYC processes involve handling a significant volume of customer data, including personal information, transaction history, and identification documents. AI technology, with its ability to process and analyze large amounts of data quickly and efficiently, can handle the volume of information involved in KYC procedures. AI algorithms can sift through vast databases, extract relevant information, and identify patterns or anomalies that may indicate potential risks or compliance issues.
Velocity: The speed at which data is generated and processed is a critical factor in KYC processes. Financial institutions need to onboard new customers swiftly while ensuring compliance with regulatory requirements. AI technologies, such as machine learning and automation, can accelerate the KYC process by automating manual tasks and performing real-time analysis. This enables faster customer identification, risk assessment, and decision-making, reducing onboarding delays and improving operational efficiency.
Variety: KYC data comes in various formats and from diverse sources, including structured and unstructured data, documents, social media profiles, and public records. AI technology can handle this variety of data by employing natural language processing (NLP), image recognition, and data integration techniques. NLP allows AI systems to extract information from unstructured text, while image recognition can process identity documents or photos. By integrating and analyzing data from multiple sources, AI can provide a more comprehensive view of customers, enabling better risk assessment and fraud detection.
How to use AI in KYC?
The process operates in a sequential manner.
Initially, it identifies distinct customers within the system, and the system examines similar ID proofs to ensure consistency with past practices. The responsible individual overseeing the KYC process then cross-references the various ID proofs associated with the customer. If the system detects customers with expired identification documents, a dubious background, or suspicious history, it assigns a relevant score based on observed inconsistencies. This information is then reported for review and further investigation.
While a lot of financial institutions are already working on automating the KYC process, some solutions may also provide an end-to-end process of digitization using application of OCR, computer vision, and basic machine learning-driven algorithms. Service providers may also tailor-made services such as AI-driven effective financial transaction monitoring, machine learning-driven alert optimization and applying AI in optimizing investigation due to the needs and request by the customers.
How many types of KYC process can be adapted to machine learning?
In this digital era, technology has come to the rescue of compliance professionals!
Starting from the launch of e-KYC where the information is seamlessly collected from different IDs available to the financial institutions through direct integration with government portals. It means that once a customer has authorized the financial institution to collect the information on his/her behalf, the financial institution, automatically pulls and fills the requisite information in their forms. Documentation requirements are reduced as the information is digitally collected from a reliable source.
So, what could be the actual function to go digital? Regarding to the latest research (Gupta, A., Dwivedi, D. N., & Shah, J. ,2023), here comes three of the key functions:
Automated Alerts for Expiry and Renewals - Based on the information updates, systems can be entrusted to generate auto alerts for the relevant staff and the customer that the documents are up for renewals, and they can be sent.
Automation of Information Extraction - This function can be achieved through scanning the e-mails, sorting, triaging, and extracting the relevant information. Human intervention can be in terms of quality control on such information collection. This technology application can save millions of hours across FIs. The important aspect of this type of solution is the accuracy of algorithms, as poorly written algorithms can spoil the trust of such systems and can also significantly increase human interventions in such processes.
Computer Vision Application on e-KYC - Financial institutes are now able to investigate the consistency of uploaded documents, e.g., whether the National ID mentioned in the form by the frontline staff is the same as the one in the document uploaded, etc., through computer vision. Discrepancies can be alerted and then investigated.
In conclusion, while present-day KYC practices suffer from loopholes and vulnerabilities. Adopting technology and sophisticated algorithms on KYC process can take away big pain, cost, and customer inconvenience in these processes. Embracing AI in the KYC process is no longer a luxury but a necessity in the evolving landscape of financial services. To know more about the automated KYC solution? Come and tell us about your need!