YOUR TEAM
Every merge stays with your analysts.
Audit trail
Every status change and human decision is captured in a chronological feed. Every auto-accepted merge creates an immutable record of the evidence, rationale, and actions taken.
Duplicate detection on key identifiers such as national ID and date of birth
Grouping into Entity Groups with similarity scores and match evidence
This could also translate into proportionally fewer alerts and cases requiring review.
One person across mortgage, card and current account, resolved into a single screened profile
A customer's full risk history — flags, true positives, case decisions — returned on any query across source systems
Why Choose The Entity Resolution Agent
A complete product, not a matching engine — review UI, workflow, adjudication and audit trail out of the box; nothing to build or integrate.
Purpose-built for financial crime compliance — entity resolution designed for the screening problem, native to the Detector, Adjudicator, Case Manager and RDM customer profile lifecycle.
Explainable by design — matching runs on national ID and date of birth, so every match traces to a specific shared identifier, never a probabilistic guess.
Human-in-the-loop — the agent does the volume, your analysts decide the merges; nothing is a black box.
Nothing lost, fully auditable — every original record preserved, every merge reversible, every auto-accept backed by an immutable decision record.
Inside Iris 7 from Silent Eight, trusted by leading financial institutions for AI-powered screening, adjudication and case decisioning.
Integrations & Deployment
Built into Risk Data Manager. The Entity Resolution Agent is an additional layer on RDM
Built on Iris 7 Deploy alongside your existing Silent Eight solutions.
Deployment Silent Eight AWS Cloud or your own cloud.
How does the agent decide two records are the same person?
It matches on existing identifiers such as national ID and date of birth, with nationality used as a conflict guard. It doesn't rely on name matching or fuzzy text scoring, so every match traces to a specific shared identifier.
Does the agent merge records automatically?
Only where you configure a safe rule to allow it — typically an exact unique-identifier match. Everything else is grouped and surfaced for analyst review. Name and date of birth alone are never enough to auto-accept.
What happens to the original records after a merge?
Every original record is preserved and retained indefinitely. Any merge can be fully unwound, restoring the pre-merge state, with the reversal recorded on the audit trail.
Is it auditable?
Yes. Every status transition and human decision is captured in a chronological group feed, and every auto-accepted merge writes an immutable decision record — inputs, grouping policy and version, evidence and rationale — at the time of the merge.
How does it run at scale?
On a delta basis — only new and changed records are processed each run, so it runs continuously across millions of profiles rather than re-scanning the whole base.
