AI-Driven Post Trade Reconciliation

Business Challenge

nova IQ was engaged as a trusted deep learning partner by a financial services firm in the post-trade space. The goal was to build an AI solution to predict which counterparty is correct during a set of trade reconciliation breaks between two counterparties for a particular securities lending contract

Solution

  • nova IQ designed and developed a CNN deep neural network and API in collaboration with the client to predict which counterparty is correct during a set of trade reconciliation breaks between two counterparties for a particular contract in 10 weeks.
  • The AI solution achieved an impressive level of accuracy (~92%)

Technologies and frameworks:

  • Python, Tensorflow, Keras, Angular, Flask
  • dAIquiri®: nova IQ proprietary framework

Impact & Benefits

  • Fewer trade breaks
  • Reduced age of trade breaks
  • Improved billing
  • Billing reconciliation
  • Improved “stickiness” for the client platform
  • User behavior changes
  • Operations personnel taking more action in the platform as opposed to offline