Automate submission ingestion and processing
Maximize revenue, reduce risk and drive operational efficiencies
Automatically extract data from broker emails and documents with high accuracy and efficiency using Eigen's best-in-class AI capabilities and reduce your submission processing time by +75%.
Get started with Submission Triage
Eigen's intuitive platform enables you to process broker submissions using one machine learning model, not many. See how easy it is to get started by requesting a free demo.
The Benefits of Underwriter Assistant
Underwriter Assistant provides flexible efficiency solutions, like Submission Triage, based on your needs, workflow and existing systems. With Underwriter Assistant you can:
Increase your submission-to-quote ratio. We'll help you spend more time on the risks that matter and less time on the risks that don't.
Increase your quote-to-bind ratio. Improve service levels to broker partners to increase binder likelihood and overall preferred underwriter status.
Reduce your loss ratio. Risk selection improves, correlation to loss is more easily understood and corrections are more effectively implemented.
Reduce your expense ratio. Improved service levels can be uses as a level to push down increased broker commission level requests.
How Submission Triage works
Process all broker submissions using one machine learning model, eliminating the need to create and manage rules or separate models when you want to extract information from individual broker submissions.
Submission capture
Add documents to Eigen via a direct email integration, API integration, S3 bucket or SFTP service.
As part of the process, the email and its attachments are uploaded for document type identification (classification) and extraction. As part of the file upload, metadata from the email such as “From Email Address”, “To Email Address”, “Email Subject Line” and “Date Received” are automatically ingested for document processing.
Add documents to Eigen via a direct email integration, API integration, S3 bucket or SFTP service.
As part of the process, the email and its attachments are uploaded for document type identification (classification) and extraction. As part of the file upload, metadata from the email such as “From Email Address”, “To Email Address”, “Email Subject Line” and “Date Received” are automatically ingested for document processing.
Classification and document association
Emails and their attachments are:
- Digitized, if necessary, to generate a text layer any application can read
- Deblobbed, to split concatenated documents into individual files so they can be processed and stored separately
- Automatically classified into their respective genres/subgenres using machine learning
- Associated with a family ID that is added to the email body and any accompanying attachments
Emails and their attachments are:
- Digitized, if necessary, to generate a text layer any application can read
- Deblobbed, to split concatenated documents into individual files so they can be processed and stored separately
- Automatically classified into their respective genres/subgenres using machine learning
- Associated with a family ID that is added to the email body and any accompanying attachments
Data extraction
Eigen understands the context around an answer, as well as the content of the answer itself, allowing it to extract the contextually correct instance of a particular term/phrase/text string from subsequent documents. By taking both into consideration, Eigen develops a conceptual understanding of the information and leverages that to extract similar information from documents it has not seen before, even if the same concept is worded differently.
Eigen understands the context around an answer, as well as the content of the answer itself, allowing it to extract the contextually correct instance of a particular term/phrase/text string from subsequent documents. By taking both into consideration, Eigen develops a conceptual understanding of the information and leverages that to extract similar information from documents it has not seen before, even if the same concept is worded differently.
Data post-processing
Customer-defined normalization and extraction rules can be applied post-extraction to standardize, clean, cross-reference or to compare data, among other things.
Customer-defined normalization and extraction rules can be applied post-extraction to standardize, clean, cross-reference or to compare data, among other things.
Data validation
Eigen provides human-in-the-loop capabilities with the ability to review the original document and remediate the extracted data. Additionally:
- Rules can be applied to any given extraction to validate data accuracy
- Each extraction is flagged as high or low confidence based on configurable numerical thresholds
- Users can leverage a manual radiation workflow to correct machine predictions, add comments or flag extractions for further review.
All changes are captured in audit logs with UI and API export functionality, and changes can also be used to improve your model.
Eigen provides human-in-the-loop capabilities with the ability to review the original document and remediate the extracted data. Additionally:
- Rules can be applied to any given extraction to validate data accuracy
- Each extraction is flagged as high or low confidence based on configurable numerical thresholds
- Users can leverage a manual radiation workflow to correct machine predictions, add comments or flag extractions for further review.
All changes are captured in audit logs with UI and API export functionality, and changes can also be used to improve your model.
Data export and integration
Data can be exported from Eigen in csv or xls format, or you can integrate Eigen directly with a variety of systems and Underwriter's Workbenches (like Salesforce, Unqork, Xceptor, etc.) using Eigen's REST API.
Data can be exported from Eigen in csv or xls format, or you can integrate Eigen directly with a variety of systems and Underwriter's Workbenches (like Salesforce, Unqork, Xceptor, etc.) using Eigen's REST API.
Data accuracy and model risk management
We take data accuracy and model risk management (MRM) seriously. As one of our core principles, we've developed our MRM framework to meet the needs of our enterprise clients. Eigen's three lines of defense ensure the data you receive and act upon is correct:
Awards and certifications
We've been the fortunate recipients of several accolades and awards as well as certifications over recent years, including the following.
Demo Underwriter Assistant
Fill in this form to request a demo of Underwriter Assistant, including Submission Triage, for your organization.
Alternatively, you can call us: