EXL introduced Medical IDP on AWS to automate the review of medical records in insurance
EXL, in collaboration with AWS, introduced Medical IDP, a solution that combines document processing with the domain model EXL Insurance LLM to automate the review of medical records in insurance. According to the company, it reduces manual review taking over 100 minutes per case to hours.
According to EXL and Amazon Web Services, insurance claims adjusters spend an average of over 100 minutes per case manually reviewing medical records. The records tend to be unstructured, often run to hundreds of pages and include dozens of document types (chiropractic reports, diagnostic tests, operative reports, psychiatric evaluations, medical assessments and more). According to the source, manual review is slow and prone to inconsistency between reviewers, leading to delayed claims payments, assessment errors, higher claims costs and increased regulatory scrutiny.
EXL, a provider of data analytics, AI and digital solutions that, according to its own figures, has been operating for over 25 years and employs more than 50 000 people, responded to this problem with Medical IDP, a solution built on the AWS platform. The solution combines automated document processing (intelligent document processing) with the domain-tuned EXL Insurance LLM model, designed to extract, summarize and query medical information at scale. Amazon SageMaker AI is used to train and run the model, while Amazon Bedrock provides access to general foundation models for broader language tasks. The solution runs within a single AWS region with access controls through AWS IAM, which the source says is important when working with protected health information.
According to the company, the EXL Insurance LLM model was fine-tuned using PEFT with LoRA on data from nine years of insurance claims operations, specifically on more than 13 500 records combining structured and unstructured data. EXL states that although general models such as GPT-4 or Claude have broad language understanding, they lack specialized vocabulary and knowledge of insurance claims workflows, which it says simple prompting cannot consistently replace. According to EXL, the solution shortens medical documentation review time from days to hours, while human oversight remains in place at key steps in the process (human-in-the-loop).
The available text of the source article also describes the technical architecture of the pipeline from document intake to structured output, but this section was not available in full. Details can be found in the source article.
Why it matters
The solution demonstrates a concrete business deployment combining a domain-tuned language model with general foundation models for a document-intensive process where errors and delays have a direct financial and regulatory impact. For insurers and similar process-driven companies, it provides an example of an architecture (fine-tuning on SageMaker AI + general models through Bedrock, human oversight of key steps) that can be compared with their own needs for automating the processing of specialist documentation.
Two audiences, two different impacts
What this means
For individuals
For individual insurance professionals (claims adjusters, underwriters), it suggests that AI with human oversight may take over some manual review of medical documentation in the future, but according to the description, decision-making itself remains with humans.
For a business
For insurers and companies processing medical documentation, this is a concrete example of how a fine-tuned domain LLM combined with general foundation models on AWS (SageMaker AI + Bedrock) can, according to the vendor, shorten manual record reviews from days to hours and reduce inconsistency and errors in claims adjudication and underwriting.
ProcessesCheck the original
Event sources
only one source so far · 1 publisher, 0 independent. We count feeds from the same owner only once.