As trade finance becomes increasingly digital, much of the discussion around trust has understandably concentrated on whether an electronic record is authentic, whether its integrity can be established and whether it has been transferred through a trusted environment.

Recent events in the receivables finance market suggest that there may be another question which deserves at least as much attention: whether the transaction evidenced by those records actually took place.

The developing dispute involving commodity trader Radiant World provides a particularly topical illustration. Proceedings before the Singapore High Court record allegations by Mizuho concerning receivables purchased in connection with purported sales of iron ore to Glencore.

According to the judgment, subsequent direct enquiries with Glencore led to confirmation that contracts presented to Mizuho did not correspond with Glencore's records and that goods had not been received under the associated bills of lading. Radiant World has disputed the allegations and maintains that the underlying trades were genuine, so the wider issues remain the subject of ongoing proceedings.

What makes the case interesting for trade finance is therefore not an assumption as to what ultimately happened, but what it tells us about the limits of documentary verification.

We are becoming considerably better at establishing whether an electronic record is what it purports to be. Digital signatures, platform controls and increasingly sophisticated verification techniques can provide levels of assurance that were difficult to contemplate when trade depended entirely upon paper. Yet none of these necessarily establishes that the commercial event described within the record occurred.

That distinction becomes particularly important as artificial intelligence assumes a greater role in documentary examination. An AI system may become exceptionally capable of comparing an invoice with a contract or transport record, identifying inconsistencies and determining whether information appears compatible across a presentation.

If the records themselves describe a transaction that did not occur, however, greater accuracy in comparing them does not resolve the underlying problem.

Perhaps this suggests that the next stage of digital trade finance needs to think about trust somewhat differently. The question is no longer confined to whether a document or data record can be authenticated, but increasingly extends to whether relevant information can be verified closer to its source and, where appropriate, independently of the party seeking finance.

None of this diminishes the value of digitalisation. If anything, it reinforces it, because connected data may ultimately allow verification that paper documents could never provide.

The difficulty is that, however confidently we can establish the integrity of a record, this does not necessarily tell us whether the trade behind it actually took place.

As trade finance becomes better at trusting digital information, we may therefore need to become equally good at understanding exactly what that trust proves.

Source: Mizuho Bank, Ltd v Radiant World Corp Pte Ltd [2026] SGHC 200

This article represents the views of the author and not necessarily those of ICC