An AI platform to understand financial documents

AI startup nRoad has introduced its Convus Natural Language Processing (NLP) platform, purpose-built for financial services with deep, domain-centric, machine learning models.

nRoad claims the platform requires minimum training samples, enables faster deployment, and is built on a microservices-based architecture that can integrate with existing IT infrastructure in a non-intrusive way while maintaining required data security.

Unstructured data, the deep, dark data that's prevalent across the enterprise but not always transparent or usable, continues to be a top business challenge. Unstructured data includes everything from documents, to images, to video and audio streams, to social media posts. Collectively, by most estimates, these types of data account for 80 to 90 percent or more of the overall digital data universe.

As unstructured data volumes keep spiralling out of control and the complexity grows, manual-heavy Robotic Process Automation (RPA) requiring extensive human interaction and generic, one-size-fits-all NLP solutions are no longer viable. The nRoad team has been on a quest to solve this problem, and has developed solutions to extract, understand, and provide insights from unstructured data.

Currently serving the financial services industry, the Convus platform:

  • Allows financial institutions to extract, normalize, and incorporate critical business information buried in unstructured documents into mission-critical business processes.
  • Offers a purpose-built enterprise-grade platform that delivers scale, accuracy, and efficiency with minimal training burden.
  • Reduces costs and avoids manual data extraction and entry; deep learning models adapt to changing document formats and structure.

"Misconceptions about how unstructured data automation can be done have persisted for decades," says Aashish Mehta, CEO at nRoad.

"This is a deep and domain heavy function that requires an understanding of what data is being presented, why it's in a certain format, and how users want to consume that data. While in stealth mode, we carefully analyzed and identified some of the key challenges and developed our proprietary Convus platform."