Mariusz Szyszka

Whether you are applying for a credit card or opening a new bank account, there is a KYC behind the onboarding process to deter criminal elements from utilizing financial institutions for money laundering activities. To kick things off, you need to know who you’re dealing with by collecting customer and business information, but are you sure you are collecting data for the right entity?

Identifiers such as trade register number, International Securities Identification Number, VAT number, etc. play a crucial role in uniquely identifying entities but does it mean identifier-based search approach ensures efficient results?

Although the identifier-based search reduces the risk of false positives associated with name-based searched there are couple of challenges to consider:

• Entities without unique identifier. For instance, in Germany, the incorporation of companies is handled by district courts, as opposed to a central register, therefore the identifiers issued are not unique across the entire country. For example, register number HRB 14567 according to handelsregister can represent several entities.
• Prefixes and suffixes assigned by data vendors to identifiers. For instance, a 9-digit identifier is assigned to every registered business in France, while some data vendors add a prefix in form of 4 extra characters to an identifier. E.g. 552018020 vs. FR38552018020
• Incorrect mapping of identifiers. It happens that the trade register number in one source matches with … the VAT tax number in another source which prevents data aggregation.
• Entities without a common identifier. Global data vendors assign unique identifiers to a single business entity (e.g. DUNS or BVD9) and sometimes it is the only identifier available for a given entity. Since they are developed and managed by a vendor, they are unknown to KYC analyst prior to conducting the onboarding.
• Entities with multiple the same type of identifiers. It has been observed that sometimes entity has assigned multiple identifiers of the same type. For instance, gleif source refers to two ISIN codes for the same entity. E.g. ISIN codes US90130A3095 and US90130A4085 are assigned to the same entity.

In addition to the above challenges, identifier-based search does not allow for a broader search, capturing variations in names or potential misspellings, therefore this approach is uncommonly used for Customer Due Diligence (#CDD) and  Know Your Customer (#KYC) in various industries.

In summary, a balanced and flexible approach that leverages both name and identifier-based searches is often recommended for AML/KYC purposes for a more robust and accurate compliance process. The key is to implement a system and algorithms that account for variations, ensure accuracy, and meet regulatory compliance requirements utilizing a combination of both methods e.g. search by name and filter by identifier and leverage deep learning and natural language processing technics for data extract and reconciliation.