Text analytics is still a relatively little known activity inside data management. However, it is a technology that is increasingly being applied inside organizations. To explain how this technology works, we’re going to explore some of the most typical and understandable text analytics case studies.
In this first group of text analytics case studies we’ll look at how text analytics increases the return on existing investments:
Drive better search
Search is a typical task in any knowledge intensive organization. In the last 20 years, companies have invested significantly in implementing enterprise search or document management systems that were supposed to help users find what they were looking for. Inevitably, however, most organizations are not satisfied with the results of traditional or even advanced search platforms. Because it enables users to intelligently filter the results of search, text analytics is probably the single highest return investment for organizations that want to address the challenge of information access. While this use of text analytics may seem obvious, it is still not a common application of the technology in organizations.
Fuel richer, higher performing predictive models
Even sophisticated models used to anticipate events that impact operations or financials are mainly based on structured data extracted from DBs or, in general, data-based enterprise applications. Quarterly sales forecasting, customer churn rates or new monthly customers tend to be heavily focused on past structured data extracted from ERPs, BIs, etc. Other information that could improve predictive models is frequently ignored because it is difficult to obtain, and even more so because it’s in a form that is different from what people are used to. In one example, a customer used text analytics to fuel predictive models that anticipate how a given board member of a public company might vote on topics that could impact the stock price. Using annual reports, they applied text analytics to automatically extract the board member’s voting performance at the given company, as well as those from any other boards (past and present) of which they were a member.
Enable more effective compliance
For most organizations, compliance traditionally means implementing workflows and procedures in accordance with internal policies or external legislation. However, being compliant increasingly requires that the form and the actual information produced are also aligned. An increasingly common text analytics case study involves the constant and frequent automatic analysis of the output of these information intensive processes (i.e. contracts, invoices, etc.) to ensure that both the workflow and the information contained within the documents is compliant.
A second area of text analytics case studies centers around automation:
Automate information intensive business processes
There are several processes that fall into this category. For publishers and media, for example, content creation, management and publication can all be more efficient thanks to automatic tagging and categorization that relieves editors from this manual task. Also for banks and insurance companies, processes such as loan application management or claim management are a typical text analytics case study because the automatic extraction of data from semi-structured or unstructured documents increases overall automation, providing considerable savings.
Enable BOTS and self help in general
BOTS, in addition to being the latest Silicon Valley craze, are set to transform customer service. Through BOTS or other natural language question answering systems on websites or other communication platforms, organizations can be more efficient and sometimes even perform better than human-based custmer support. This is a typical text analytics case study because the high performance is directly linked to the ability to automatically understand the meaning of words, which is at the core of semantic-based text analytics software.
Keep DBs current
Several enterprise systems suffer from not being updated with the latest information. Take HR systems, for example:Employee data, especially regarding skills, are not updated unless employees go through formal training sessions, which is increasingly uncommon in many organizations. What about on-the-job skills that employees acquire over time? These are also rarely communicated with HR. Text analytics can be very effective for ensuring that DB information is kept in line with the current reality. In this case for example, it is enough to constantly analyze the content produced by employees, as well as their official performance reports or promotions within the organization.
