Digitization Studio
Digitization Center enables you to configure and manage schemas across all lines of business using natural language. Configuration users define file classifications, fields and source mappings using natural language, unlocking fast, scalable deployments and enabling insurers, brokers and reinsurers to author their unique view of risk. Configuration users can take advantage of out of the box schemas accelerating time to value and reusability across different transaction types and lines of business. Schemas can be iteratively tested prior to production using a simple test interface. The Cytora platform makes it easy to add new fields to progressively evolve and expand your view of risk.
Time to value
Scalability
FAQs
To create a new classification you need to provide a natural language name of the document and a description. Up to three levels of hierarchy can be created to nest classifications within higher level classifications. No coding required
To create a new schema field you need to provide a natural language name for the field, the description of the field and the output data structure. No coding required.
Yes it is. After creating classifications, configuration users can define source mappings which control which documents fields come from.
Under field types users can choose the type of field you want to create ranging from numbers to text to repeating groups and dropdowns. For example, nested arrays can be created by using repeating fields (e.g. properties) each have fields (e.g contents) associated with them. This enables users to achieve a digitization output which corresponds with their target data structure.
Cytora has a guided experience in which configuration users can test schema performance against samples of submissions. Users are provided with explanations of why classifications were chosen and can see field candidates extracted relative to the target schema. This allows users to iterate schema design to achieve the best level of performance
Yes they can. Users can copy existing schemas into new digitization schemas as well as copy individual fields. This enables reusability across schemas especially with fields that are common across different transaction types and lines of business.
The Cytora platform allows confidence levels to be assigned to both file classifications and fields. Users can define confidence thresholds which, if not met, activate human review tasks
Digitizers work immediately after being configured in natural language, requiring no labelling of data or training runs, unlike legacy machine learning solutions, returning near-instant time-to-value. Our ‘Teach’ interface allows you to test your configured digitizers against sample submissions to gauge their performance and tweak where required.
What our customers say
“
Cytora will enable us to grow our scale across our specialist markets, embed and reinforce a data-driven culture and empower our underwriters to provide better service to customers and brokers.”
Commercial Director
Contact form