by Stephanie Dalwin

Something all startups have in common is a love of data, regardless of the solution they provide. Increasingly, startups are lending their expertise with unstructured data. Startups don’t usually provide pure data sets or data aggregation, as many large and established vendors have a foothold in these spaces. Yet many offer platforms to ingest data, process it, and derive insights.

Digitizing handwritten forms and unstructured data reading and ingestion is a growing space, especially in property/casualty insurance uses cases like streamlining submission and intake., for example, applies NLP and AI to unstructured data sources (e.g., insurance documents) for use cases like data extraction, policy checking, and submission triage and prioritization. CogniSure built an insurance-specific platform that extracts unstructured data from loss runs, submissions, quotes, and policies, among other documents.

Unstructured data ingestion is also helping streamline underwriting. Groundspeed Analytics uses machine learning to extract data from insurance products to enable frictionless underwriting and streamline intake. Convr (formerly DataCubes) uses machine learning to digitize data from policy documents like applications and loss runs and uses third-party data to derive insights and automatically answer insurer underwriting questions.

Beyond underwriting, InsureTech startups are solidifying their expertise in using unstructured data for claims predictions and insights. SD Refinery, for example, uses AI to consume and process unstructured sentence data from document files or database tables for use in areas like claims and underwriting. Clara analytics applies AI and machine learning to extract structured and unstructured data from medical notes, bills, and other documents to generate predictions like loss costs and propensity for litigation.

Insurers produce vast stores of data, but often it is locked in unstructured documents. The market for unstructured data ingestion and insight solutions is growing, as is the number of startups that tackle insurance documents specifically. These vertical solutions may help insurers bring unstructured data analysis in-house without having to commit too many internal resources.

You’ll find more information on startups tackling unstructured data in our InsureTech industry report, InsureTech for Insurers: 200 Startup Profiles. Novarica will also be releasing an update in September that contains 250 profiles.

As always, feel free to get in touch with me or Novarica EVP Jeff Goldberg if you’re interested in discussing InsureTech trends and the current landscape. You can reach me at, and you can reach Jeff at

This article first ran in