Identify · Attack atlas

Eight equal-priority hypotheses, not a ranked threat list.

The atlas translates emerging GenAI-enabled fraud into 8 testable campaign archetypes and 32 variants spanning cards, account-to-account payments, identity, merchants and disputes.

Illustrative detailcard-not-present

Adaptive card-testing swarm

Validate credentials before monetisation.

GenAI enabler

Generative agents vary low-value purchase semantics and timing at scale.

Campaign envelope

$1–$85

6–18 linked events

State sequence

1probe
2cooldown
3retry
4monetise

Scenario variants

micro-authorisation probesmerchant-hop testinglow-and-slow retrieswallet-token validation

Mitigation hypotheses

velocity controls
merchant-card graph risk
step-up authentication

Validating artifact schemas and preparing the command center…

Eight equal-priority hypotheses, not a ranked threat list.

Adaptive card-testing swarm

AI-personalised APP scam into mule network

Deepfake-assisted account takeover

Synthetic-identity credit bust-out

Agentic-commerce intent hijack

GenAI fake-merchant collusion ring

AI-crafted invoice and QR redirection

Synthetic-evidence refund abuse

Adaptive card-testing swarm

Eight equal-priority hypotheses, not a ranked threat list.

Adaptive card-testing swarm

AI-personalised APP scam into mule network

Deepfake-assisted account takeover

Synthetic-identity credit bust-out

Agentic-commerce intent hijack

GenAI fake-merchant collusion ring

AI-crafted invoice and QR redirection

Synthetic-evidence refund abuse

Adaptive card-testing swarm