Should I use a biometric KYC system?

DTI 'LSS Suite': Money Laundering Act extends KYC principle; new tools accelerate the examination

In view of the money laundering laws of the individual nations, the international data sources and the tougher penalties that are due when favoring money laundering or terrorist financing, obliged companies quickly run the risk of paying a horrific penalty if their KYC checks are not up to date to organize. With the LSS Suite from DTI (website), this task should be accomplished in a fraction of the time - an astonishing number of conversion and normalization tools create a comparable database.

[With the new Money Laundering Act in Switzerland, notaries, brokers and large digital platforms with crypto currencies such as Bitcoin and Ethereum, art brokers are now obliged to carry out KYC risk analyzes (KYC - Know Your Customer) if they have sales contracts with a volume of over 10,000 euros Process cash payments. In the meantime, the group of “obligated parties” goes far beyond the financial sector. This includes goods companies, so in the future, for example, car dealerships, luxury goods and precious metal dealers, who also have to check the origin and ownership of the cash from 2,000 euros, similar to gambling providers.

Checks and corrections for text and speech recognition with LSS Suite
With the text and speech recognition of content to be indexed, controls and corrections are necessary for the security of KYC checks. The LSS Suite uses the following basic functions and strengthens its individual effects with matching algorithms:
Tokenizer break down the text of a search entry into a sequence of logically related units, into so-called tokens, with which more refined searches are possible in a text.
Spell check checks the correct spelling when entering a search; in the case of a "wrong" or misspelled term, it provides possible suggestions for correction (did-you-mean).
Fuzzy matching delivers results even with incorrect and alternative spelling (typing errors, twisted letters, text recognition errors).
Phonetic matching delivers results with similar sounding terms (e.g. Meyer, Maier, Meier)
Phrasing recognizes standing terms such as "First August". This ensures that both terms must appear in close proximity and are not spread out somewhere in the text.
Anti-phrasing removes words that are not or less relevant for the search. When searching for a company, e.g. additions such as AG, Ltd, GmbH, etc.
Normalize normalizes and transforms search entries in order to support various forms of representation, such as a date or amount.
Lemmatizing is the function with which a search term is reduced to its basic form or expanded to the various forms. Thus, the term "house" also finds variants such as "houses" or "house".
Synonyms extends the search to similar or identical names; for example using alias lists. This means that when you search for Peter, you can also find Petrov or Petrovic. This is often used in addition to mathematical algorithms such as phonetic (sounding similarly) or fuzzy search (similarly written).
Thesaurus are often branch-specific dictionaries in order to take specific terms into account when searching; For example, bank-specific terms such as shares, securities, funds or organizational structures.

Extensive duties of care and documentation

KYC requires a clear identification and verification of new customers as well as a regular follow-up analysis of existing customers and a documentation of these risk assessments. The companies obliged to carry out the KYC analysis are thus part of their authorities' prevention strategy in order to identify and prevent money laundering and terrorist financing at an early stage. Identifying a contractual partner also includes establishing the actual ownership and the beneficial owner (s) and checking the economic background. For the onboarding of a new customer and the regular due diligence reviews of existing customers and their payment flows, this means that dozens of internal and external data sources have to be evaluated each time, each consisting of several million data records. An evaluation of the internal data sources with the available search functions of relational databases may still succeed quickly; but a comprehensive examination of external sources is only possible with an IT solution at an economically justifiable expense.

The solution: Name matching with ListSearch Compliance Server

Name matching is at the center of the IT-based KYC analysis tool from DTI Schweiz AG, the ListSearch Compliance Server (LSS Suite).

In order to find the person or company you are looking for from a variety of data sources in a risk analysis, complex algorithms for semantic, linguistic and phonetic analyzes are used for the search and identification process in order to first record and index database contents. The problem here is that the data from the various sources is only partially structured and unstructured data such as continuous texts, video or audio files as well as packaged or protected files such as ZIP or PDF cannot be indexed for subsequent automatic evaluation without preparation. Often, text recognition programs first have to read documents and audio recordings have to be converted into text (speech to text) before they are even available for analysis. And errors occur when evaluating using OCR (Optical Character Recognition) or speech-to-text conversion.

In addition, there are different writing conventions, people, place and company names are written in different variants. In addition, there are rotated numbers, umlauts or character sets that are unknown in many languages.

What even search engines like Google or Bing sometimes do not succeed in doing has not been taken for granted with search engines in corporate applications, but it is indispensable for KYC. "

In contrast to the functions of databases, which can largely only deal with structured information and mostly only provide Boolean operators (and, or, nor) for searching, the LSS Suite search solution is able to use standardized interfaces to retrieve all digital information and prepare and index data. With the LSS Suite, the focus is on the connection of all digital data sources available in the company such as file systems, customer relationship management and enterprise content management systems, scanned and OCR recognized documents, voice and video files.

The LSS Suite would be able to integrate, evaluate and index different types of data sources via connectors. So that customers do not have to give their own sensitive customer data out of hand, the LSS Suite is an on-premise solution that runs stationary on its own servers. External data sources are read in once and updates are reloaded daily as required. Connectors have the task of opening up databases, file system sources (World-Check, Thomson Reuters, SWIFT) but also web content (press, PEP and lists such as OFAC). For the data sources, the LSS connectors are also capable of incremental indexing, i.e. the automatic indexing of data changes. For external sources, it is also a decisive advantage that the LSS Suite is based on an open architecture such as XML or ReST (Representational State Transfer).

The data collected via the connectors are processed by the LSS Suite before being transferred to the indexer. Filters (conversion) are used to convert the different file formats (Word, Excel, PowerPoint, PDF, HTML, RTF, TEXT, ZIP) into a uniform, readable format for the indexer.

In addition, checks and corrections are necessary for the speech recognition of the content to be indexed. In addition to the usual functions such as tokenization, synonymous representation, spell checking, lemmatization, entity extraction, phonetic, etc., the LSS Suite provides additional, highly specialized functions for name matching (see box).

Key functions in the indexer determine success and security

The indexer in the LSS Suite forms the core of the search functions and ultimately the success of the system as well as its ability to identify the correct matches and automatically sound an alarm if a hit is found. The task of the indexer is to structure the data internally during indexing in such a way that results are available within seconds, even with extremely large amounts of data and with complex and very specific searches. Based on this index, it is also possible to actively inform the user automatically about new content that meets the predefined criteria. That is why the LSS Suite provides further query options in addition to a simple query language like Google.

DTI Schweiz AG has been an independent system integrator and expert in the fields of data analysis, text recognition and processing as well as data management and search for over 25 years. It integrates individual, highly flexible solutions based on the latest technologies and products. A development team takes care of the individual needs of the customer. ListSearch Compliance Server (LSS Suite) from DTI Schweiz AG offers all components for a successful name matching workflow from internal and external data sources. The LSS Suite consists of a web interface for operation, a search service with application logic and a specialized search engine. The LSS Suite offers the possibility of automatically processing lists with data and having them searched for using the solution. Thanks to configurable, flexible search templates, the various options of the LSS Suite can be used almost without restriction for any matching and search scenarios. The search service prepares the results, evaluates them and automatically assigns them to the responsible processors, and the search results for the data records are documented for complete traceability.
The name-matching framework of the LSS Suite masters various questioning techniques and puts them together in a targeted workflow into so-called search cascades in order to cover various aspects of the name search. The user does not have to deal with the formulation of complex questions, but can only enter the criteria to be searched for. The LSS Suite processes the rest fully automatically. The output of the results is prepared for the user in such a way that he only has to worry about the evaluation of the hits identified as potentially relevant. The ranking of the result output is also a quality factor of the LSS Suite. Because only with a good ranking is the "wheat from the chaff", that is, relevant from irrelevant hits, separated. Depending on the interest group in the company or the type of search application, the requirements for the ranking of the results can vary. That is why the LSS Suite also offers the option of adapting the ranking to the various needs of the users.

Return on investment in the shortest possible time and documentation of the KYC analyzes

For users with a few dozen checks a day, the tool pays for itself in a very short time, according to the manufacturer. Even experienced compliance employees would need an average of at least five minutes for each customer. This effort is also regularly due with existing customers. Even at medium-sized banks, with up to 70,000 new and existing customers per year, the result is quickly three to four full-time positions that can be substituted using LSS Suite. aj

You can find this article on the Internet on the website:

(No ratings yet)
Loading ...

Interesting too