John Hancock is an insurance company offering a broad range of products including life insurance, long-term care, annuities, and 401(k)s. Here is how ANTstein Square from AntWorks helped them achieve greater than 92% accuracy for handwritten text as well as a few other benefits.
The company receives large volumes of documents that require manual review and processing. John Hancock must classify the documents and extract the unstructured data, primarily handwritten text and signatures. Once extracted the data is then passed downstream for the execution of the client request in their admin systems. Client requests include a variety of forms such as auto-debit for insurance premiums, withdrawal requests, address changes, and change of beneficiary forms.
Antwork’s intelligent automation platform, ANTstein Square, ingests a variety of different documents and data formats and feeds them into end-to-end automation. Forms are analyzed across different processes using machine learning and natural language processing to make data-driven decisions. Using ANTstein square’s cognitive machine-reading and machine learning capabilities, the quality of the documents is first enhanced with the pre-processing engine, then they are classified using the document indexing engine. Handwritten text is extracted dynamically from the documents and reinforced using advanced deep learning techniques.
ANTstein square helps John Hancock achieve greater than 92% accuracy for handwritten text. Furthermore, accuracy increases with more samples in training:
- 75% reduction in manual data extraction;
- increased productivity;
- decreased processing time;
- easier workflow management.