The traditional claims handling process was highly manual. For the initial claims reporting, a client or broker submits a claim by phone or email, triggering a potentially lengthy process to gather all relevant details. This includes collecting documents and images detailing the damage, estimates on repairs, any injury data where applicable, and so on. Often an on-site evaluation is required to assess and determine the validity of the claim, generating potentially hand-written adjuster notes. At some point, a fraud detection process is also initiated, again a highly manual process.
For many years carriers have been trying to automate claims processing by using optical carrier recognition (OCR) technology to help process documents by converting hand-written and printed text to a machine-readable format. While an improvement, OCR required templates to read most documents, making it vulnerable to errors when a given piece of data was not in the exact spot the template expected it to be. That created a costly process of near-constant template maintenance to keep up with new or changed documents. Given that, along with the need for manual review for accuracy, OCR could hardly be considered a solution for insurance claims automation.
Intelligent intake represents a quantum leap forward in insurance claims automation. A form of intelligent document processing, intelligent intake obviates the need for templates. Instead, it uses artificial intelligence technologies, including large language models, natural language processing, machine learning, and transfer learning. Taken together, these technologies enable insurance companies to build intelligent intake models capable of ingesting nearly any kind of document or image and extracting relevant data from it. That data is then converted into a structured format that can be entered into downstream claims processing systems, such as Guidewire Claim Center.
The ability to ingest unstructured documents, extract relevant info from client claims and turn it into a structured format represents a significant step in insurance claim process automation, and gives rise to numerous important capabilities and use cases.