What Is Trade-Based Money Laundering and How Can AI Help Banks Detect It?

Answer-first summary

Trade-based money laundering, or TBML, uses trade transactions to disguise criminal proceeds or move illicit value.

AI can support detection by comparing data across documents, parties, goods, prices, shipment information and transaction activity to identify inconsistencies and unusual patterns.

What is trade-based money laundering?

FATF defines TBML as disguising the proceeds of crime and moving value through trade transactions in an attempt to legitimise their illicit origins.

The underlying shipment or transaction may appear legitimate. The manipulation can be hidden in the value, quantity, description, route, documentation or financial structure of the trade.

Common TBML techniques

Potential techniques include:

1. Over-invoicing goods.
2. Under-invoicing goods.
3. Issuing multiple invoices for one shipment.
4. Misrepresenting the quantity or quality of goods.
5. Shipping more or fewer goods than declared.
6. Using false or altered documentation.
7. Providing vague or misleading commodity descriptions.
8. Routing payments through unrelated third parties.
9. Using complex company structures without a clear commercial purpose.
10. Moving goods through unnecessary jurisdictions.
11. Reusing documentation across different transactions.

No single indicator proves that money laundering has occurred.

A red flag identifies an issue that may require additional information, investigation or escalation.

The four main groups of TBML indicators

FATF groups TBML risk indicators into four principal categories.

1. Business structure indicators

These relate to the organisation behind the transaction.

Examples include:

1. Business activity that does not match the goods being traded.
2. Unverifiable operating addresses.
3. Complex ownership structures without a clear commercial purpose.
4. Individuals who appear to act as nominees for beneficial owners.
5. Transaction volumes inconsistent with the organisation’s apparent size.
6. Sudden activity following a long period of dormancy.

2. Trade activity indicators

These concern the commercial logic and behaviour of the transaction.

Examples include:

1. Unexpected trading routes.
2. Unnecessary intermediaries.
3. Excessively complex transaction structures.
4. Goods inconsistent with the customer’s line of business.
5. Transactions that appear uneconomic.
6. Sudden high-value activity by a newly established business.

3. Trade document and commodity indicators

These concern the documents and goods being presented.

Examples include:

1. Different prices across contracts and invoices.
2. Conflicting quantities, weights or values.
3. Vague goods descriptions.
4. Missing or potentially counterfeit documents.
5. Frequent amendments without clear justification.
6. Goods routed through several jurisdictions without commercial reason.
7. Prices that appear inconsistent with market value.

4. Account and transaction indicators

These relate to the financial behaviour associated with the trade.

Examples include:

1. Payments from unrelated third parties.
2. Last-minute changes to payment instructions.
3. Account activity inconsistent with the customer’s stated business.
4. Rapid movement of funds through several accounts.
5. Transactions deliberately kept below reporting thresholds.
6. Short periods of unusually high activity followed by dormancy.

Why is TBML difficult to identify manually?

The information required to assess a transaction is often fragmented across documents, internal systems and external data providers.

A reviewer may need to compare:

1. Customer and counterparty details.
2. Prices and quantities.
3. Goods descriptions.
4. Shipment routes.
5. Port information.
6. Vessel data.
7. Previous customer activity.
8. Screening results.
9. Payment instructions.
10. Internal risk rules.

One data point may appear reasonable in isolation. Its significance becomes visible only when it is connected with the rest of the transaction.

How TraydGuard supports TBML detection

TraydGuard analyses trade-document data and presents potential risk indicators in the context of the wider transaction.

Its capabilities include:

1. Automated TBML red-flag checks.
2. Cross-document comparison.
3. Sanctions and AML intelligence.
4. Vessel and shipment checks.
5. Pricing-anomaly identification.
6. Severity-based risk scoring.
7. Traceable findings.
8. Institution-specific rules.

The objective is not to determine automatically that a transaction involves criminal activity.

It is to help specialists identify which transactions warrant additional investigation.

“Compliance is no longer just a checkpoint—it’s a continuous, data-driven process. Agentic AI allows banks to treat compliance proactively rather than reactively, significantly reducing risk and improving transparency with regulators.”

Tarun Rishi, CPTO, Traydstream

Strengthen your TBML controls with intelligent trade-document and transaction analysis.

Explore TraydGuard.