Loading FusionStore...
PartnersResourcesSupportDeveloper Portal

Fincom.co's AML screening

FINCOM.CO

Payments
Compliance Screening
Add-on module to enhance your current AML screening system
Payments
Compliance Screening
Add-on module to enhance your current AML screening system

Description

Fincom.co’s Real-Time AML compliance solution uses purpose-built, supervised Machine Learning with “Phonetic Fingerprint” technology. It allows for screening high volumes of transactions at low latency, from multiple languages and industries, while dramatically reducing false positives without missing hits. 

In a world where financial transactions are becoming immediate and demand real time verifications, manual processes under AML compliance consume enormous resources that result in extremely high operational costs. Moreover, US AML regulations are becoming even more strict: lower amounts, more and more industries must comply, fines are getting bigger and personal liability is now official. Fincom.co’s Real-Time AML compliance solution aims at addressing those 3 main challenges in ALM today.

Real-time low latency screening

AML verifications using rules-based/libraries solutions take between 3 seconds to 3 minutes on average. Fincom’s solution speeds up the process by evaluating the transactions against the largest Sanction/PeP databases, performing an accurate, precise screening under 200 milliseconds, while still ensuring "No Misses".

Greater accuracy while reducing false positives

Financial institutions hire thousands of trained expensive HR, costing up to 10% of the entire operational costs on financial crime only to screen false positives. By using a supervised machine learning technology, the false positive rate can be reduced by over 50% from the industry standard​.

Multilingual “no miss” by Phonetic Fingerprint

Financial Institutions are fined due to outdated technology that can’t keep up with the diversity & volume of industries & languages. Using a mathematical representation of the way the spelled name sounds, this enables matching names at far greater accuracy, eliminating the problem of spelling variations with over 38 pre-built languages​.

Company Verified

General information

Commercial model
Resell
Integration
Pre-integrated products
Fusion Global Pay Plus
doneAvailable for all 3rd party systems
Implementation
Cloud Hosted | On Premise
Website
Support
Data processing countries
Israel

How it works

Fincom.co’s technology solution screens any name of a person or entity against the largest Sanction/PeP databases such as Dow Jones Factiva, Refinitiv World-Check, that accumulated over 8.5 million entries with associated 170 million names and perform an accurate precise screening under 200 milliseconds. The two kinds of filtering thresholds, the phonetic threshold that checks name composition and the name-distance Threshold that resolves the difference between names that spelled or sounds alike are both adjustable. The alerts are automatically resolved based on past approvals data, using ML trend analysis.

Onboarding Customer

The customer information “Phonetic Fingerprints” of Entity is compared in wide funnel (multi-lingual) – Prevents Missed-Hits (False Negative), whilst AI/ML engines are applied to filter out False Positives.

Potential Alerts

Alerts are transferred upwards to the Pending DB which contains ‘flagged’ customers names.

Resolution

By manual verification, the name goes into the “Green All Clear Customer Database”.

Ongoing Verification Process

Every day, the entire customer roster is verified against Changes in Sanction & PEP lists, to ensure no customer is sanctioned.

Getting the results

Results are pushed by a standard Rest-API into the user known UI/UX.
How it works

How it looks

Welcome Screen

Real-Time matching of any Entity by Name Pronunciation or “Name-Sound”, based on Single Mathematical Representation.

Search result for queries

Using the most advanced solution for name matching to date, which incorporates Computational Linguistics.

Raw results

The platform allows the user to access the results as row data.

Filtered results

Filters can also be applied in order to identify specific results.
screenshot container
Welcome ScreenSearch result for queriesRaw resultsFiltered results

Welcome Screen

Real-Time matching of any Entity by Name Pronunciation or “Name-Sound”, based on Single Mathematical Representation.

Resources