signzy

API Marketplace

downArrow
Logo
Responsive
Decorative line

Identity Graph

Overview

An identity graph links identifiers (emails, phones, devices, cookies, addresses, government IDs) to a unique person across channels and time. In compliance and fraud, it powers duplicate detection, householding, and detection of multi-accounting or mule networks. Nodes represent people or entities; edges capture relationships (shared device, address, funding account).
Graph-based features degree, centrality, community feed risk models and casework. Enrichment brings in registry checks, sanctions hits, and geolocation to contextualize connections. Governance is critical: define permissible purposes, minimize PII, and log lineage. With a resilient identity graph, institutions reduce false positives, spot coordinated abuse, and tailor due diligence by network exposure, not only by individual records.

FAQ

Why use a graph at all?

Relational patterns expose rings and duplicate accounts that flat tables miss. Connections reveal hidden risk and speed investigations with context.

How is it built safely?

Start with strong identifiers, add probabilistic links with thresholds, and keep provenance. Limit access and retention to regulatory expectations.

How does it help KYC?

It flags duplicates, links related accounts, and prioritizes EDD for customers connected to high-risk nodes or jurisdictions.