Data collection has been at the heart of insurance business processes since the birth of the industry. As insurers advance into the digital age, they need to reevaluate how they access their data, especially the rich information they have stored on paper forms. In particular, it is becoming imperative for them to move from expensive and time-consuming manual data capture to automated processes.
Many steps are needed before the rich rewards of big data can be reaped; of these, choosing the right data capture method for an insurance organization is one of the most important. Best-fit technology yields the best results. Tools and methodologies insurers might consider as part of their data capture solution include:
Word recognition technologies
Three primary technologies are commonly used to convert images or PDFs of static documents containing typed or handwritten text into machine-encoded text: Optical character recognition (OCR), intelligent word recognition (IWR) and intelligent character recognition (ICR). Some of these technologies have existed for many years, and have been consistently improved over that time.
In order to achieve high accuracy rates, companies must spend a significant amount of effort and resources configuring and optimizing the technologies for each and every type of form to which the technology is applied. Even then, a significant amount of manual quality review and data entry capacity is required.
Application of text recognition technologies
Recognition technology is often integrated with document management systems (DMS) and more broadly, enterprise content management (ECM) systems. ECMs have enabled certain global insurance firms to overhaul their entire legacy system and business information communications throughout every level of their enterprise.
While ECMs can improve a company’s IT infrastructure and increase efficiency, they are seldom the best single solution for data capture. Implementation can be expensive and time-consuming. However, today’s leading-edge technology providers can equip insurers with reliable, cost-effective data capture and transformation software that achieves extremely high accuracy levels and requires relatively little integration.
Mobile data capture
With mobile data capture (MDC), users take images of documents using a camera-equipped mobile device. The image data is sent to a data capture server that extracts the information stored in the image.
MDC is flexible, scalable, and increases speed while significantly decreasing the operating costs associated with data capture. It can be used effectively as the platform of a digital-first capture strategy; data is entered on the device and moves directly into a back-end system.
Crowdsourcing and machine learning
The merging of human and artificial intelligence is an approach to data capture that is at once flexible and scalable. Software alone, while far-reaching in its abilities to capture information from static sources, cannot accurately contextualize data. That ability remains a function of human intelligence. Crowdsourcing can thus be used to significantly enhance data capture software.
Data as a service
Data-as-a-service (DaaS) solutions such as those from Captricity make captured data available in a record that can be viewed and downloaded from a secure website or imported directly into existing CRM and ERP systems and statistical analysis tools. All interactions with the product occur via a web browser, agent, or application process interface (API).
The right approach: Man + machine
The right data capture approach can help insurers create value by reducing back-office costs. The automation of manual data entry tasks (e.g., forms processing) has the potential, in addition to decreasing costs, of increasing operational efficiencies and improving response times to customer queries.
A man + machine approach can quickly, inexpensively and accurately provide the structured data needed for analytics-driven insights, which in turn are critical to success in today’s marketplace. Such an approach combines advanced machine computing and analytics with the superior contextual recognition powers of the human brain to accurately recognize printed and handwritten text and transform it to structured digital data.