Articles

Articles

Optimize the utility of Unstructured Text with Text Analytics

Sandeep Arora - EVP & Head - Enterprise Data Management

Widely used for conducting Market Research, Text Analytics is a contemporary technology and a subset of Advanced Analytics. Also referred to as Text Mining among highly technical groups, it is the process of excavating quality intelligence from largely unstructured text by identifying patterns and trends. In a way, it is a process of deriving meaning out of the free text generated during customer engagement through emails, online query resolutions, surveys, etc.

Business Intelligence, the precursor of Business Analytics and now its subset, is centered heavily around analysis of organized or structured data. Times have changed. Today, the data that is generated and stored is increasingly in unstructured form. The potential to dig out valuable intelligence even from free text, gives Text Analytics the proverbial status of ‘philosopher’s stone’, which is capable of converting base metal to gold. This technology proves immensely helpful for market research as well as decision making. Though the process is excruciating, the extra miles put in are worth the effort.

Text Analytics provides manifold benefits, such as capital management (especially in BFSI sector), exploration of new risks, fraud detection and tracking, etc., and thus enables enterprises to carry out informed decision making. It can also be used to provide highly proficient and customized services to customers. Customer communication analyses as well as customer-specific customized products have tremendous potential for business growth.

Enterprises generate huge volumes of unstructured data. The realization that mining of unstructured data has overwhelming value has triggered business lines to culture, grow, and harvest it. Following are some immediate uses that generate value through Text Analytics:

  • Customer Services: Text Analytics has high potential value in Customer Services sectors. The tons of structured and unstructured data that is generated during customer engagement can be wisely gleaned to extract information, such as key customer concerns, preferences, etc. Retaining customers is an uphill task. Hence using Text Analytics as an adjunct for customer profiling ensures key issues are addressed immediately.

  • Research and Development: High volumes of text are gathered from research papers, competition analysis, industry-specific information, market research, etc. This information proves to be a rich source for digging valuable intelligence for generating market leads, market trends, new product generation, etc.

  • Fraud Detection: Fraudsters find unique ways to trick the systems. Text Analytics lends immense value to detect fraud patterns and apprehend the fraud cases. Patterns spotted in the text of insurance claims give out the details about possible fraudsters. Text Analytics thus shields insurers’ systems from expensive remodels and upgrades of technology to counter fraud cases.

  • Risk Management: Health Insurance, Vehicle Insurance, and Property Insurance claim settlements, which always show a particular pattern in the claims at large, can be used to assuage risky practices and fine-tune underwriting. For example: Areas with high saline content in the soil can damage the foundations of buildings and areas near to water bodies could dampen walls causing mould that weakens the construction. Such cases can be effectively dealt with proper premium assignment when high risk geographical areas are detected and slotted by using Text Analytics while monitoring insurance claims.

  • Loss Reserving: Claim settlement is a complex process and at times requires a long time to settle when claims are escalated to regulatory authorities and courts. Using Text Analytics, insurers can expedite processes, quickly pick up patterns, verify for authenticity, and check precedents well in advance for building up their loss reserves. The insurers can also quickly come up with pricing structures based on forecasted levels of risk and also reduce the financial impact of claim escalations. Also through effective usage of Predictive Analytics and Text Analytics, insurers can predict well in advance the reserves that need to be assigned for claim settlement for a specified time period and produce a figure of underwriting income for statutory compliances in a timely manner.

  • Capital Management: In the course of customer dialogue, a huge source of ideas, concerns, and content is generated. The enterprise needs to address these concerns or voices-of-the-customers at the earliest. Using Text Analytics, the executives can assess and assign priority to these concerns and maintain the operational liquidity accordingly in order to fulfill the requirements of their customers.

 

Text Analytics helps enterprises extract meaningful information from the unstructured text that is generated during business processes and customer engagement. These large reservoirs of unstructured text is a gold mine for extracting information that proves immensely helpful in areas, such as Customer Services, Research and Development, Fraud Detection, Risk Management, Loss Reserving, Capital Management, etc. The value thus generated through Text Analytics is enormous and has great potential for business growth.