ezrules blog
The ezrules blog
Practical writeups on building, testing, and shipping fraud, risk, and compliance rules with ezrules.
-
How fraudsters prepare their attack infrastructure
Learn how fraudsters prepare for a fraud attack and what you can do to spot it earlier.
Read post → -
Graph-Based Event Investigation in ezrules
ezrules adds graph-backed event investigation and graph-derived rule features, revealing shared cards, devices, and identities to catch fraud rings and clusters.
Read post → -
From Zero to Live Rule Evaluation in 10 Minutes with ezrules
Spin up ezrules and prove the full fraud-rule workflow in 10 minutes: author a rule in the UI, evaluate a live event via the API, and confirm the decision.
Read post → -
Rule Lifecycle Management in ezrules: Draft, Promote, Pause, Roll Back, Archive with Audit Trail
ezrules adds an explicit rule lifecycle: draft, active, paused, archived, with auditable promotion, reversible pause, and rollback that restores known-good logic.
Read post → -
AI-Assisted Rule Authoring in ezrules: Generate Transaction Monitoring Rules from Natural Language
Turn plain-English intent into validated, reviewable transaction-monitoring rules with ezrules AI authoring, complete with diffs, explanations, and human-in-the-loop control.
Read post → -
Shipping Rule Changes with Less Guesswork: Shadow Deployment in ezrules
Shadow deployment in ezrules evaluates candidate rules against live traffic in parallel with production, so you validate real-world impact before promoting a change.
Read post → -
Rule Quality in ezrules: Precision and Recall by Outcome-Label Pair
The ezrules Rule Quality view scores every rule on precision, recall, and F1 by outcome-label pair, surfacing over-flagging and missed cases against ground truth.
Read post → -
Ordered Rule Execution in ezrules: When First Match Beats Conflict Resolution
ezrules adds ordered execution and a first_match mode for the main rule lane, letting teams stop on the first matching rule with full audit trail and governance.
Read post → -
Automatic Field Type Management
ezrules normalizes inconsistent JSON field types at the engine level, so rule comparisons stay correct without per-rule casting, with a full audit trail of changes.
Read post → -
Why Complex Entities Matter in a Fraud Rules Engine
ezrules supports nested dotted entity paths across rule logic, testing, observations, and backtesting, so fraud events model real customer, sender, and device relationships.
Read post →