Moroccan customs authorities have used AI and digital risk-analysis tools to audit imports worth nearly 650 million dirhams ($65 million), recovering close to 50 million dirhams ($5 million) in duties, fines and related charges after finding declared values inconsistent with market prices.
The account comes from sources familiar with the matter speaking to Hespress AR. No customs official has been named and no official statement has been published.
How the system works
The National Customs Investigation Unit analysed large volumes of import data before isolating files showing discrepancies between invoice prices and reference values for the same or similar goods in countries of origin.
The specific pattern it looks for is repeated imports of similar goods at significantly different prices — the signature of systematic undervaluation rather than isolated error.
Officials then compared invoices and commercial documents against:
- Internal customs databases
- Data exchanged with foreign customs authorities
- Information held by professional organisations
- Prices recorded among foreign suppliers and in external markets
That is a more specific methodology than most government AI announcements describe, and it is the kind of problem the technology genuinely suits: finding anomalies across datasets too large to review manually.
Where it led
The investigation extended to invoices from foreign exporters flagged with elevated risk indicators, several of them based in China and other Asian countries — an output of the risk scoring rather than a finding about any particular jurisdiction.
Officials examined the commercial relationships between those exporters and Moroccan importers, along with the sequence of invoices, shipping operations and payments.
The audits also reached sectors that had not previously been priority targets for customs monitoring, after data analysis identified risks in declared values.
Linking customs to tax records
A further element is worth noting on its own terms.
Customs has strengthened investigations by cross-referencing its data with information held by other public administrations and institutions, comparing import declarations against companies’ commercial, tax and financial records.
That is effective for enforcement, and it is also a substantial expansion of state data integration. An importer’s customs filings, tax returns and financial records becoming jointly queryable is a capability with implications beyond undervaluation — and one that would normally attract data governance scrutiny.
Why undervaluation matters beyond revenue
Suspected undervaluation affects not only customs receipts but competition in the domestic market, according to the sources, by reducing the tax and customs costs borne by some importers relative to compliant competitors.
That is the argument for enforcement that does not depend on the revenue figure: importers declaring honestly are disadvantaged against those who do not.
Based on reporting by Hespress AR.
