Parascript unveils FormXtra.AI 8.0 for insurance and financial services

Parascript has included new machine learning capabilities in FormXtra.AI 8.0, the latest version of its Intelligent Document Processing (IDP) platform.

Using new deep learning recognition (in addition to transcribing large blocks of handwritten information such as those in comments fields of all types of forms), Version 8.0 adds the power of NLP to transcribe entire pages of handwritten information - reading similar to a human with results that demonstrate the best performance in the industry.

Previous releases used results of page-level classification to derive document boundaries. With 8.0, separation is improved by using deep learning neural networks to review multiple page features including image, temporal context and page sections to identify logical document boundaries. The novel application of multiple areas of context means FormXtra.AI can automatically separate files into documents even where separation of two documents is required within a single page.

Complex semi-structured and unstructured documents often have important data that exists as a block or paragraph of text making it difficult to locate and extract all of the needed information without erroneously missing some portions of the text or including other unnecessary sections. With FormXtra.AI 8.0, a new sentence and paragraph context is added that simplifies these situations.

Using only labels such as “address,” the surrounding areas can be analysed, and the entire address data can be located or all the data within specific paragraphs based on section titles can be located and extracted.

Parascript has added the ability for Smart Learning to automatically detect if a field is mostly static on a form or if it is randomly placed as in a semi-structured or unstructured document, lending much more adaptability and high-fidelity support of the variations of complex document-based information.

For those who want to add their own rules in addition to the auto-generated ones, FormXtra.AI 8.0 provides a user experience that makes blending the results of deep learning classifiers with additional rules a simple process.

In practice, these expanded capabilities have significant impact on complex IDP workflows. Examples of this can be seen in both mortgage automation and in healthcare insurance automation.

Loan Processing Optimization

Mortgage lenders can significantly accelerate their loan onboarding process. Using Parascript software, they can immediately validate the documents submitted by borrowers and ensure that a complete loan package complies with any number of stacking orders. The ability to verify the presence or absence of all the mandatory documents saves downstream process time.

In addition, lenders have a new automation option so that they can easily extract all the necessary information from these packages, validate the data and pass it downstream to all their applications without delays. 

FormXtra.AI significantly reduces document processing times with high accuracy — simultaneously reducing the cost of quality control and avoiding the errors introduced by manual processes so that more documents can be processed each day.

Insurance: Moving Toward Interoperability

Using Parascript Smart Learning, FormXtra.AI can be trained on any document from medical charts to claims. The software identifies key characteristics of each individual medical record.

Different algorithms are often employed that evaluate various attributes such as presence of graphical information (e.g., logos), textual data (e.g., facility names and addresses), and even spatial information such as the distance between different dates on a page and use of specific language related to those dates.

All these attributes are then analysed to identify the most reliable way to identify and separate one document from another. Once separated, data extraction can then be employed with the final step being to employ Intelligent Capture to further identify specific patterns in the text.

In addition to accurately accessing and rapidly moving data through the enterprise systems, this data can reveal various problems and conditions not only with a single patient, but across a patient population.