Duplicate detection

Spotting that a new applicant is already in your system.

For example, someone who was declined under one name might apply again under a different name, and duplicate detection recognizes that the face or the details match an existing record.

Deduplication of enrollments via biometric or attribute-matching against existing records.

Duplicate detection works in a few steps:

1. The system collects the new applicant's face, details, and device information.
2. It compares them with existing enrollments.
3. It scores how closely any existing record matches.
4. A strong match blocks the application or sends it for review.

The comparison can draw on several kinds of data:

• Biometric matching, which searches enrolled faces for the same person.
• Attribute matching, which compares details such as name, date of birth, phone number, email, address, and document number, often with fuzzy logic that tolerates typos and spelling variants.
• Device and network signals, which show that several applications came from the same device or connection.

Duplicate detection sits next to terms that are easy to mix up:

• Face search (1:N) is the biometric technique that compares one face against many, and duplicate detection often relies on it.
• Identity graph links accounts through shared attributes to reveal clusters, which complements one-to-one duplicate checks.
• Synthetic identity fraud builds identities from real and invented details, which duplicate detection can expose when the same face or device appears under several names.