THE TECHNOLOGY BEHIND
Breaking New Ground in AML Sanctions Screening
Phonetics at the Core
Fincom’s innovative approach provides transparent, consistent, accurate, and explainable results, eliminating the need to rely on outdated rule-based or obscure AI technologies.
INNOVATIVE APPROACH TO SANCTIONS SCREENING
Fincom approaches the challenges of Sanctions Screening from a totally different angle, using any written representation of the sound of a name as the fundamental “building block,” as opposed to traditional approaches that rely on how the name is spelled, e.i., written (e.g. Token, N-Grams, etc.). In combination with 48 proprietary algorithms in phonetics, computational linguistics, and advanced mathematics, this approach leads to outstanding screening accuracy and speed.
Fincom’s innovative approach provides transparent, consistent, accurate, and explainable results, eliminating the need to rely on outdated rule-based or obscure AI technologies.
PHONETIC FINGERPRINT
A single mathematical representation of phonemes (the Name Sound) within a name (individual or other entity)
Phonetic Fingerprint is Fincom’s patented technology based on Phonetics. It allows building a universal common denominator to create a solid background to compare the sound of written names in 40+ languages against a target word (name or other entity) in real time.
Phonetic Fingerprint technology allows “translating” any Name – Person, Company or Entity – into a mathematical sequence that represents the way the writing sounds, overcoming different transliterations, name structures and spelling variations, achieving accurate matching of names across and between 40+ different languages in real-time.
Creating a common denominator – matching numbers to numbers
Accurate – Efficient – Effective
PHONETIC-LINGUISTIC ENGINE
The core technology in Fincom’s products for AML & Entity Resolution.
Utilizing phonetics, computational linguistics, and mathematics to transform an entire name structure into a mathematical sequence – Phonetic Fingerprint ID to then use the Phonemix Formulae to measure distance between phonetic fingerprints, ensuring Accurate Name Matching.
Exceptional ability to measure the true level of Phonetic Proximity.
OVERCOMING LANGUAGE BARRIERS
40+ LANGUAGES
in original Alphabets
Overcoming Language boundaries:
Fincom’s phonetic-linguistic engine based on the proprietary Phonetic Fingerprint technology allows for names screening and matching within and across languages, recognizing both names in their language of origin as well as name transliterations. Unique ability to “understand” and match unstructured names in a multilingual environment. Currently the language pool includes over 40 languages in original Alphabets among them: Russian, Chinese, Arabic, Spanish, Korean, and others.
PHONETIC-LINGUISTIC
ENGINE
based on PHONETIC FINGERPRINT
proprietary patented technology
MULTI-LAYERED DATA
PROCCESSING
for cross-database analytics
in 40+ languages & diverse data formats
SMART SUGGESTION
ENGINE
based on PHONETIC FINGERPRINT
proprietary patented technology
Fincom's core technology allows to:
IDENTIFY
names of individuals or companies in multiple languages in original spelling in 40+ languages
ENSURE
AML compliance + accurate results that guarantee effectiveness and high efficiency.
PREVENT
name-based AML discrimination, resolving complex & unstructured names
REDUCE
alert rate and operational burden by over 90%
SEARCH&MATCH
entries, including transliterated & unstructured names with spelling mistakes & variations
MATCHING ENTRIES
ACROSS MULTIPLE
DATABASES
Capable of matching and resolving differences in record fields for each record by creating a common cross-language & cross databases denominator for DB entries unification. Creating strong entity unique identifiers by resolving and combining a Name (individual, company, or other entity) with an additional field (e.g. Address, IBAN, or other), while each one of the identifying components may be derived from different data sources.
The technology allows for:
- Cross-database analytics: anonymized data search & sharing throughout multiple data silos and diverse data formats.
- Analyzing data in 40+ different languages in original Alphabets or transliterations.
- Effectively cross-reference data of different formats across multiple data silos.
INSIGHTS
Finding a needle in a haysteck: addressing the challenges of aml name screening
Several factors contribute to name variations, including competing indexes, initials, abbreviations, middle names, nicknames, standardized names as per IGI, letter substitutes, and translations across languages. These variations add to the complexity of name screening, making it a formidable challenge in combating criminal activities such as money laundering and terrorist financing. With global business transactions and internet banking on the rise, AML screening must encompass the understanding of all the commonly used languages, each with its own unique alphabet, pronunciation, and diverse forms.