With the advent of the internet, the world witnessed information presence anytime and anywhere. People no longer needed to be at a library skimming pages to seek information. Instead, most of the information is right there at arm’s length.

On the other hand, hasn’t it happened to all of us that at times this availability of a vast amount of information gets too overwhelming for us to even process, at times leaving us confused, and worried about the uncertainty or reliability of the information? 

Information overload is a real problem in today’s world. And like with many of the modern-day man-made problems, we have come up with various names for it, such as infobesity, infoxication, information anxiety, and information explosion. This situation becomes more severe in sensitive areas of finance, management, healthcare, where even a minor mistake can cause repercussions of catastrophic magnitude. However, as necessity is the mother of invention, there are many smart minds tirelessly working to solve these challenges, especially in the financial service domains- where a single bad decision is all you need to wipe away someone's life savings. 

Capital Quant Solutions (CapQuant), is focused on solving this challenge by extracting structured data from unstructured financial documents. The company offers tools that leverage the power of machine learning and NLP to identify relevant parts of the document and analyze it. 

According to Capital Quant “the financial services sector is inundated with unstructured data and this deluge of information should be converted to strength for the industry and not a weakness because of overload.”

The company is the brainchild of Pravin Lal, who is its Founder & CEO. Pravin has over two decades of experience in banking and capital markets tech with mainstream organizations such as Citi Bank, Lehman Brothers, Nomura Holdings, Headstrong Inc, and Sapient. 

IDC estimates that an average knowledge worker in the financial services industry spends around 2.5 hours per day searching for information from unstructured sources. This, according to per IDC calculation, translates to a cost of 33 million USD per year for Goldman Sachs alone. 

Currently, this process is done manually, where an analyst reads and scans the documents, extracts relevant information, and then analyzes the same. Such analysts are hired across all functions such as credit, risk, trade finance, equity, fixed income, etc. Manual process has many limitations such as being expensive as it consumes expensive human bandwidth, they are often error-prone, and are not scalable. Today banks and financial institutions across the globe receive a high volume of complex and unstructured documents such as financial statements, news, corporate announcements, and research reports. As a result, many of the key financial decisions are made based on the information in these documents raising high demand for tools such as FinStinct. According to Research and Markets, the Global Business Process Automation Market size is expected to reach $19.4 billion by 2026, rising at a market growth of 13.2% CAGR during the forecast period.

“The global financial services industry will experience major disruption in the coming decade in specific process areas which deal with decision making based on information from unstructured documents. Currently, this industry hires intelligent resources i.e. the analysts who read such complex documents and manually identify relevant details which are extracted, structured and analyzed for decision making.We believe with domain-specific AI solutions such as FinStinct from Capital Quant, such processes will experience disruption as ROI gains are significant,” says Pravin Lal, Founder & CEO.

CapQuant developed FinStinct as a comprehensive cognitive automation solution that uses the power of ML and NLP for processing just about any financial document and pulls out the relevant information for analysis. FinStinct has pre-created models which are available off the shelf for generic documents which almost all financial institutions use. It adapts to your information needs and helps you focus on more integral tasks for business growth.

For the financial service industry, many processes such as underwriting processes, investment analysis, legal processes, KYC processes, and regulatory processes require key details to be extracted from unstructured documents and then analyzed. 

As per company sources, 80% of the data in a financial services organization comes from unstructured sources, yet less than 1% of this data is analyzed and used. Working with unstructured documents requires an intelligent system that can understand the document’s structure (Document Object Model) automatically, even for previously unseen styles and layouts. It is expected to be capable of extensively using computer vision and deep learning to perform document structure analysis and extract text from irregular shapes such as seals and stamps in legal documents. Identification of relevant portions of the document (which could be textual or tabular) for a specific use case and of specific Key-Value pairs which can come from a paragraph or free-flowing text, table, or form is critical. The solution is expected to understand the structure of non-uniform tables and accurately extract it. 

While most of the solutions available in the market work on specific documents only, FinStinct can work on any document as it is a comprehensive cognitive automation solution.

Another distinguishing factor of FinStinct is that it offers very strong DIY capabilities to a business user who is new to machine learning in building an ML model to extract key data points.  

With the complex nature of the problem and lack of standardization in the way the data is represented in the documents, the detailed testing of the models becomes time-consuming. 

FinStinct is currently being used for over 25 use cases at NSE. In addition to that, over 35 different document types being processed at the exchange. At NSE, FinStinct is an integral solution for the digital transformation of the exchange. Documents from all divisions of the exchange covering regulatory compliance, clearing, collateral management, index management, member services, member compliance, listing compliance, etc., are being processed by FinStinct. Few other clients who are using the tool are IIFL, DMI Finance, and SmartStream.