For over 100 years, this Fortune 500 title insurance company has helped smooth real estate transactions by providing comprehensive title insurance protection and professional settlement services. With over 18,000 employees, this company offers its products and services in more than 40 countries.
Agents examined over 35 types of unstructured documents related to each property manually to evaluate the chain of title sanctity. There are several steps needed to generate a title report, from investigation, to checking vesting of the property, legal description and financial matters. The entire process was time-consuming and error prone. Business rules were complex and knowledge retention was poor, resulting in additional work for supervisors – reviewing sensitive information and ensuring accuracy. As a result, there was high staff turnover, and hiring and training new employees was costly. In addition, the client’s existing OCR platform struggled with the variability of document formats which required constant intervention. The sheer variety of these documents made a template-based automation approach inefficient and ineffective.
The insurance company implemented Cognitive Machine Reading (CMR), a combination of AI, cognitive technology, Machine Learning and pattern recognition. CMR accelerate the title verification process. It reads 35 unstructured document types and extracts 1,260 data points. Its pre-processing capability automatically enhances the readability of old property documents. Natural language models generate inferences and improve its ability to read and classify legal documents.
After CMR was implemented, turn-around time for work processes decreased by 80% while ensuring 100% regulatory compliance. The customer achieved a 200% increase in productivity, 360-degree view of data points across all 35 documents, and automated inferences. CMR reduces the processing time from 3 weeks to 3 days for a single case leading to redeployment of resources within other functions, reducing operating cost and improved efficiency.