KYT — Know Your Transaction¶
Definition¶
KYT (Know Your Transaction) is the process of monitoring and analyzing customer transactions to detect suspicious patterns that may indicate money laundering, terrorist financing, fraud, or other financial crimes. While KYC verifies who the customer is, KYT monitors what they're doing with their account.
KYC vs KYT¶
graph LR
subgraph "KYC — Point in Time"
A[Who is this person?]
B[Are they who they claim?]
C[What's their risk level?]
end
subgraph "KYT — Continuous"
D[What transactions are they making?]
E[Are patterns consistent with profile?]
F[Is anything suspicious?]
end
A & B & C -->|Onboarding complete| D & E & F
style D fill:#4051B5,color:#fff
What KYT Monitors¶
Transaction Red Flags¶
| Red Flag Category | Examples |
|---|---|
| Structuring (Smurfing) | Multiple deposits just below $10,000 to avoid CTR reporting |
| Unusual volume | Account with $2K/month suddenly transacts $200K |
| Geographic risk | Transfers to/from FATF grey/blacklist countries |
| Round-tripping | Money sent out and returned through different channels |
| Rapid movement | Funds deposited and immediately transferred elsewhere |
| Cash-intensive | High cash deposits inconsistent with business type |
| Third-party payments | Payments from unrelated third parties |
| Shell company patterns | High throughput, no employees, minimal business activity |
| Layering | Complex series of transactions designed to obscure origin |
Transaction Monitoring Rules¶
graph TD
A[Transaction Event] --> B[Rules Engine]
B --> C[Threshold rules<br/>Amount > $10K]
B --> D[Pattern rules<br/>Structuring detection]
B --> E[Geographic rules<br/>High-risk country]
B --> F[Behavioral rules<br/>Deviation from profile]
B --> G[Network rules<br/>Connected entity analysis]
C & D & E & F & G --> H{Alert Generated?}
H -->|Yes| I[Investigation Queue]
H -->|No| J[Normal Processing]
I --> K[Analyst Investigation]
K --> L{Suspicious?}
L -->|Yes| M[File SAR/STR]
L -->|No| N[Close Alert]
style M fill:#e53935,color:#fff
KYT in Crypto (Blockchain Analytics)¶
KYT has special importance in cryptocurrency:
| Challenge | Traditional Finance | Crypto |
|---|---|---|
| Identity | Account holder known (KYC'd) | Wallet address is pseudonymous |
| Transaction tracing | Bank records | Public blockchain ledger |
| Cross-border | Correspondent banking, SWIFT | Instant, borderless, 24/7 |
| Speed | Hours/days for settlement | Seconds/minutes |
Crypto KYT providers:
| Provider | What They Do |
|---|---|
| Chainalysis | Blockchain analytics, wallet risk scoring, transaction tracing |
| Elliptic | Cross-chain analytics, sanctions screening for crypto |
| TRM Labs | Blockchain intelligence for financial crime |
| Crystal Blockchain | Crypto transaction monitoring and compliance |
FATF Travel Rule¶
The Travel Rule requires VASPs (Virtual Asset Service Providers) to share originator and beneficiary information for crypto transfers above $1,000:
| Data Shared | Details |
|---|---|
| Originator | Name, account number, address/DOB/ID number |
| Beneficiary | Name, account number |
| Transaction | Amount, date, virtual asset type |
KYT Technology Stack¶
| Component | Technology | Purpose |
|---|---|---|
| Rules engine | Custom rules, Drools, FICO | Threshold and pattern-based detection |
| ML models | Anomaly detection, graph neural networks | Detect sophisticated patterns |
| Graph analytics | Neo4j, TigerGraph | Network analysis — connected entities |
| Real-time streaming | Kafka, Flink | Process transactions as they happen |
| Case management | NICE Actimize, Featurespace, custom | Manage investigation workflow |
Key Takeaways¶
Summary
- KYT is the ongoing monitoring complement to KYC's onboarding verification
- Monitors transactions for suspicious patterns: structuring, unusual volume, geographic risk
- Suspicious Activity Reports (SARs) must be filed when suspicious activity is confirmed
- Crypto KYT uses blockchain analytics to trace pseudonymous transactions
- The FATF Travel Rule mandates information sharing for crypto transfers
- Modern KYT uses a combination of rules and ML for detection