Fintech Friday: Nephelai – machine learning for trade booking and processing

Eléonore de Vial, co-founder and CEO at Nephelai, spoke with Efma’s Boris Plantier about how her company is transforming the way financial institutions process trades and transactions.

Fintech Friday: Nephelai – machine learning for trade booking and processing

What led to the creation of Nephelai?

I started my career as a consultant in charge of implementing middle/back office solutions, and as such was in contact with many middle office teams of investment banks and asset managers. The majority of tasks accomplished by middle office teams revolve around management of errors, either directly spending time correcting them or upstream, their identification via numerous reconciliation processes (confirmation, cash, position and trade reconciliation, valuation control, performance control, etc.).

Seeing the rise of AI and machine learning in many fields and its capacity to automate more and more complex tasks, I realized that it could be a strong candidate to help financial institutions prevent errors, thus saving time and reducing their risk.

After working over 10 years in a large financial software company, my associates and myself gathered the knowledge, the experience, and the skills to launch a startup using machine learning to detect and prevent errors on financial transactions. We were also thrilled to create a solution of our own, which could live up to our own performance and quality standards.

Could you present Nephelai's offer?

Nephelai clients are investment banks, asset managers, asset servicers, brokers, and large corporations with dealing desks.

We leverage machine learning techniques to transform the trade processing chain. We use the audit trail of past transactions to train models which are then used to analyze new transactions. More specifically, at the front office level we propose smart defaulting: the trader fills a few fields and we predict the content of many others. At the middle office level, we scan transactions in real-time, detect potential mistakes, and propose corrections. Finally, we give management a consolidated view of the quality of their operations, through a KPI dashboard aggregating error rate by desk/activity and identifying actions offering the best leverage for overall improvement.

The solution is mostly applied to financial transactions, market data, and referential instruments.

Société Générale has been using the solution in production since June 2019. Their stock loan traders use the smart defaulting function to increase the quality and speed of the booking process and the alert blotter is used by control teams to guarantee that no errors remain unattended for long.

OFI Asset Management has been using the solution since October 2018 to first identify areas to improve, second assess STP and confirmation rates, and lastly to guarantee that their overall effectiveness remains at the highest standards.

What's coming next for Nephelai?

We are working on a new way to book transactions: Instead of filling an old fashion form or excel spreadsheet with many fields, traders will be able to book trades like they do a google search on the internet. This smart google-like bar will be easily integrated into any existing capture screen to combine old-fashioned booking and a new way of booking. We are lucky enough to have customers willing to work closely with us, so that we are sure our solutions work in real life conditions

Banner_fintech.gif (16 KB)


Trading Fintech AI/Robotics


This website is not compatible with Internet explorer please use one of the following browsers: Chrome, Edge, Firefox