top of page

Paperless Trade Solution Launched By Traydstream

Traydstream is the new fintech player that has launched a digital solution to trade documentation and in so doing, aims to automate regulatory compliance screening using artificial intelligence.

“Trade is one of the few bastions of banking that remains extremely manual. We thoug

ht, there has got to be a way to automate some of this.” – Uzair Bawany (co-founder Traydstream)

Traydstream’s solution is to digitalise the whole trade transaction – from invoice to Swift – and is aimed at being useful for banks as well as corporates.

Other companies, such as Bolero and essDocs, have already developed similar platforms for paperless trade. However, the Traydstream founders believe their solution to be unique in that it processes the documents and checks them against a database of tens of thousands of global and regional trade finance regulations and rules. These processes in most banks are manual and thus labour-intensive jobs.

“Document checks tend to be done by very experienced senior professionals who have been in the bank for many years, who know all the tricks. We’ve made that into a smart process using technology. So what typically takes a human being between four and 10 hours, we’re aiming to do in three minutes.” – Uzair Bewany

The engine uses semantic analysis techniques, and can quickly check all trade data for a range of issues; blank fields, the inconsistency of names, industry-specific legislation, sanctions and country restrictions. This also helps banks to tackle anti-money laundering and compliance sensitivities.

The machine is not only more efficient and faster, but it is often more reliable, they say. And because the system uses machine learning technology, it only learns and develops as it is fed more data.

Together with a team of ex-trade finance bankers and industry professionals, Tradestream co-founders Bawany and Achille D’Antoni have spent the last two-and-a-half years developing the new solution. The company has spent US$1.2mn building the platform, self-funded by 15 shareholders, and is now looking to raise series A financing to commence pilot tests of the software.

“A lot of people who know how to read a document and how to process it are about to retire. This is a global problem. This means they have to be replaced by others, and the cost of training, to bring people to that level, is huge,” - Achille D’Antoni (co-founder Traydstream)

The first step of Traydstream’s process is to convert all trade documents to a digital format. To do this, Traydstream has developed an OCR (optical character recognition) engine, which can read, scan and instantly structure and store paper-based information digitally.

“Trade business is made up by paper with different formats, fonts, issuers and structures. We looked at a dozen OCR products, but none served the idea that we had, namely, to have an OCR that could put together all the different formats and make it work. That’s the reason we had to develop it ourselves.”- Achille D’Antoni

Their OCR solution’s accuracy on the extraction of clean data is 97%. And this data can be stored on the cloud.

However, the goal is a world where there is no paper in the first place, and thus an OCR is not needed. “We are trying to work with clients to load data straight onto our portal, so the scanning is not even required. That would be our dream. The OCR is the headache we wish we didn’t have to deal with” says Bewany. The Traydstream founders also have many ideas for how the platform could be expanded and integrated with other technologies:

“There’s a possible connection with blockchain. The digital information from the blockchain can be put onto our platform, which removes all the OCR issues, so we see blockchain as an enabler for us, which is quite exciting.”- Uzair Bewany

The founders also see the possibility of expanding the solution to a crowd funded platform where private investors could chip in to a letter of credit. Currently, the goal is to develop and test the solution as it stands though.

bottom of page