ComplyCube's AML Screening matching engine is purpose-built to find and match comparable identity data from any language or script. It incorporates sophisticated fuzzy matching algorithms, heuristics, linguistic variance, and phonetic analysis to find matches without producing costly and time-wasting lists of false positives.
To simplify usage, ComplyCube represents the outcome of the analysis in the form of a match score. A match score can range from 0 - 100, with 100 being the best match.
By default, an AML Screening Check will only return matches with a match score over 85 and automatically reject matches with a score ranging between 85 and 90. Ultimately, only matches with a score higher than 90 will require manual checks.
You can adjust these automation thresholds according to your risk-based framework and internal process.
By default, ComplyCube will not exclude any matches. We will return matches without a DOB, gender, nationality, or incorporation country details. Furthermore, ComplyCube will also return inactive, deceased, and low-quality alias matches.
Depending on your risk tolerance and preference, you can select from two distinct name matching techniques:
Fuzzy (default): applies various linguistic, phonetic, and near-match techniques on names to find matches.
Precise: only matches names that are exact
The search mode only applies to the name matching aspect of the Screening Check. ComplyCube utilizes other proprietary algorithms to find matches using other attributes.
ComplyCube's Document Check uses OCR technology and advanced analytics to authenticate and extract data from your clients' ID documents.
This setting empowers age verification compliance by allowing you to set a minimum acceptable age before your client ID is flagged.
By default, a Document Check will flag anyone younger than 18 years of age.
ComplyCube extends further optional checks, including:
Client data validation: this check compares the client record data (e.g., first and last names) with data extracted from an ID document.
By default, ComplyCube will not perform any additional checks.
ComplyCube's Identity Check utilizes computer vision, liveness, spoof-detection technologies, and facial comparison algorithms to determine whether two faces belong to the same client.
Facial similarity threshold
ComplyCube represents the outcome of a facial analysis in the form of a facial similarity score. A facial similarity score can range from 0 - 100, with 100 being an exact match.
By default, an Identity check will deem two faces a match when the facial similarity score is over 80.
ComplyCube returns a liveness score, representing the outcome of a liveness analysis. A liveness score can range from 0 - 100; the closer the score is to a 100, the greater the confidence that your client is a real person.
By default, an Identity check will deem a client real when the liveness score is over 50.
Enrolled faces limit
ComplyCube returns a previouslyenrolledface indicator, asserting whether a face was registered as a different client.
By default, an Identity check will deem a face previously enrolled when registered more than once.
Proof of Address Check
Document age threshold
ComplyCube's Proof of Address Check aims to establish the issuing date of an eligible document using OCR. If a document is older than a specified threshold (e.g., 3 months), the document age analysis will be flagged for the check.
By default, a Proof of Address check will automatically flag documents older than 3 months. You can adjust the threshold to be between 1 to 12 months.