This company turns unstructured data into actionable insights for businesses

Aashish Mehta, the Chief Executive Officer of nRoad, is on a mission to unlock the potential hidden within unstructured data across enterprises. His journey began with pivotal roles at Genpact, where he spearheaded sales and marketing for AI solutions, driving digital transformation in banking and financial services. Now, at nRoad, he leads the charge in leveraging cutting-edge AI to shed light on what's often termed 'dark data,' revolutionizing decision-making and innovation for organizations worldwide.

nRoad stands at the forefront of AI-driven innovation, with its flagship platform, Convus, leading the charge in transforming unstructured data into actionable insights. Celebrated for its traceability, continuous evolution, and robust security measures, Convus has rapidly gained trust among top-tier clients in banking, insurance, and financial services sectors. In an interaction with Bizz Buzz, Mehta sheds light on nRoad's evolution and future plans, emphasizing the critical role of AI and ML in analyzing unstructured data

What is the nature of your business and how did it evolve?

nRoad has been years in the making. We saw a need in the market for a platform to help handle the growing dilemma of unstructured data. Using our past experiences we created the Convus platform, which today is trusted by global enterprises to help optimize processing of their unstructured data and generate relevant intelligence.

What are the key focus areas for nRoad?

nRoad is out to help bring order to the chaos of unstructured data. nRoad’s main focus is helping to illuminate dark data, by creating business outcomes for enterprises from their unstructured data in an efficient, accurate and scalable way.

How does nRoad plan to continue staying at the forefront of AI-driven innovation in the enterprise software industry, and what upcoming developments can we expect from the company?

nRoad’s team has years of experience in both financial, AI and automation technologies. This experience has allowed our team to see the unique needs and increasing demands facing enterprises. nRoad’s Convus platform is able to learn and grow with companies that choose to adopt this service and the possibilities provided are limitless. You will see major releases of Convus by enabling LLM and GenAI components to further improve accuracy of the outcomes and increased interactions with unstructured content.

What are the specific achievements or milestones that nRoad has reached so far in its journey to revolutionize the processing and utility of unstructured content?

We achieved over 95 per cent extraction accuracy, 2.8 seconds faster processing time than popular cloud-based options. Being used and trusted by major global financial institutions.

What distinguishes nRoad's Convus platform from conventional NLP solutions? Can you provide a brief overview of the key AI and ML technologies that power Convus and their role in turning unstructured data into actionable insights?

Traditional NLP solutions fall short because they are unable to handle the domain specific concepts, relationships and business context that are needed when dealing with unstructured data for an enterprise. Current solutions are heavily template-driven and support static table/database data sources. The evolution to a structuring of complex sentences and narratives upon comprehension of complex input sources is highly distinctive. Multilingual documents and integration with third party applications dominate 80 per cent of enterprise data architecture. Incorporating institutional and domain knowledge for continuity and scalability is critical. Having the ability to traceback the reasoning and factual statement identification on the input source enables confidence in the system. The foundational flaw in generic approaches (RPA, OCR Software, and cloud based solutions) dealing with unstructured data is the lack of consideration in input variability, content variability, language and localization constraints. nRoad leads with a patent-pending technology framework that leverages deep learning techniques. The framework enables companies to mine and extract actionable insights from both structured and unstructured data sources.

What are the key features and benefits of Convus in terms of processing unstructured data and enhancing business operations?

We understand that enterprises are riddled with native and cloud based applications. Instead of asking an enterprise to fit into our solutions, we have designed our platform to fit within an enterprise’s existing architecture. Deploying a stateless, event-driven and containerized solution provides enterprises the freedom and flexibility to operate by their own rules.

Deep Neural Net (DNN) methods with an inherent knowledge base acting as a foundational layer for Machine Comprehension (MC), let us take unstructured data to a new level. Our advanced text analytics allow the machine to read, comprehend and retrieve contextual information that meet our customers’ requirements. Added multilingual capabilities also gives them the flexibility to scale their business processes across the globe.

What strategies does nRoad employ to keep Convus continually evolving and adaptable to the changing landscape of unstructured data and documents?

Convus has a self-serving training module embedded in it to train and retrain, allowing us to meet our customer’s specific needs with accuracy and speed. It is continuously learning. The Convus platform is also designed to function modularly. This modularization makes it possible for every piece to be retrainable and configurable, and allows for rapid onboarding of new use cases bringing unstructured data to a new level.

What are your thoughts behind the formation of nRoad. How is nRoad helping clients eliminate the unstructured data, and helping them reach their goals.

I first became aware of the world of unstructured content processing and its challenges back in 2006 when we founded Corporate Fundamentals, and it was clear then that this space was going to continue to grow. With the volume of unstructured data continuing to pile up and the complexity of this data continuing to grow, the generic, one-size-fits-all NLP solutions that businesses have relied on for years are simply no longer up to the task. Our founding team saw a massive opportunity. The foundational premise of nRoad is that understanding both the domain and the underlying business context is the essential first step in solving the unstructured data problem. nRoad’s Convus platform helps businesses looking to maximize their growth and market reputation by leveraging AI and ML solutions to extract valuable insights from their unstructured data.

What is the role of artificial intelligence and machine learning in analyzing unstructured data?

nRoad’s Convus is a modular platform focused on deconstructing unstructured data. The consumption engine is capable of decomposing any form of document to a standard structured representation or seamlessly integrating with any data source to consume information. Information is contextualized and comprehended in the Natural Language Understanding (NLU) engine, by leveraging knowledge graphs and domain knowledge. Our Generation Engine is capable of producing bespoke outputs and constructing narratives using Natural Language Generation techniques. Convus integrates within a businesses’ existing framework, so there is no outsourcing of information when using this solution. Corvus is able to analyze unstructured data at faster speeds and with higher levels of accuracy than traditional manual data analysis methods.

What are your future plans?

The problem of unstructured data is not going away anytime soon. Enterprises that want to stay ahead of the curve will continue to look for ways to manage and analyze their unstructured data. We continue to future-proof nRoad, by working on developing our own language models that are capable of reading financial documents and unstructured content with high levels of accuracy and precision. We plan to release our own language model to deal with financial statement analysis soon. Additionally we also plan to explore other domains, such as healthcare, where there are large amounts of unstructured content.