Notifications
Clear all
Topic starter
12/07/2026 12:15 pm
TL;DR: Fraud teams are being forced to redesign around AI-driven abuse, account compromise, and social engineering as transaction volume rose 18% in 2025 while payment fraud attempts held near 3.25%, according to Sift. The operational issue is no longer only detection speed, but whether the team structure, metrics, and escalation model can absorb synthetic abuse at scale.
NHIMG editorial — based on content published by Sift: The Operational Blueprint for Modern Fraud Teams
By the numbers:
- Transaction volume on the Sift Global Data Network grew 18% in 2025, while payment fraud attempt rates held steady at around 3.25%.
- Dispute rates rose 78% year-over-year in Q3 2024.
Questions worth separating out
Q: What breaks when fraud teams stay focused only on payment loss reduction?
A: They miss account takeover, onboarding abuse, synthetic identities, and social engineering that damage trust without always producing an immediate chargeback.
Q: How do fraud and IAM teams work together on AI-driven abuse?
A: They should share the same decision points.
Q: What do security teams get wrong about trust and safety org design?
A: They often treat it as a staffing choice instead of a control model.
Practitioner guidance
- Reclassify abuse by attack surface Map your fraud cases across payments, account takeover, onboarding abuse, synthetic identity, support fraud, and internal social engineering before you rework staffing or tooling.
- Align fraud and IAM escalation paths Create a shared escalation model for account compromise, step-up failures, suspicious identity proofing, and privileged internal requests so security and fraud teams are not operating separate triage queues.
- Replace loss-only reporting with outcome metrics Track approval rates, false positives, customer churn after fraud events, and manual-review throughput alongside loss rate so leadership sees the business cost of overblocking and underblocking.
What's in the full article
Sift's full post covers the operational detail this post intentionally leaves for the source:
- How Kevin Lee structures fraud, trust, and safety organisations across centralised, embedded, and CoE models
- The full metric shift discussion on approval rates, false positives, churn, and business-outcome reporting
- Career and hiring heuristics for fraud analysts, including how to evaluate inquisitiveness and proactive investigation
- The operational roadmap for the next session in the series, including detection, response workflows, and review design
👉 Read Sift's operational blueprint for modern fraud team design →
AI-driven fraud pressure: what trust and safety teams need to change?
Explore further
12/07/2026 12:29 pm
AI-scaled fraud has become an identity governance problem, not just a loss problem. The article shows that fraudsters are now targeting accounts, people, and internal decision points with AI-assisted speed. That shifts the centre of gravity from blocking bad transactions to governing identity assurance, review authority, and exception handling across the business. For identity teams, the fraud function now depends on the same assurance controls that IAM and verification programmes manage every day.
A question worth separating out:
Q: What should organisations measure if they want to know fraud controls are working?
A: Organisations should measure whether controls are increasing attacker cost, reducing campaign success rates, and forcing repeated abuse to become uneconomic. A control can reduce one attempt and still fail strategically if attackers can immediately retry at low cost. The right metric is not only detection, but deterrence.
👉 Read our full editorial: AI-generated fraud pressure is exposing trust and safety org design gaps
