← Back
AWS ML Blog

Reducing medical claims review time with AI on AWS: The EXL Medical IDP solution

•15 min read•
#llm#deployment
Level:Advanced
For:AI Engineers
✦TL;DR

EXL’s AI‑powered Medical Intelligent Document Processing (IDP) solution reduces the > 100‑minute manual review time for insurance claims by combining a template‑agnostic IDP engine (Xtrakto.AI) with a domain‑specific large language model (EXL Insurance LLM) on Amazon SageMaker. The pipeline ingests, splits, classifies, extracts, enriches, and post‑processes heterogeneous medical records, then provides summarization, natural‑language querying, and traceable reasoning over ICD and CPT codes. This end‑to‑end AWS‑hosted architecture enables scalable, secure production inference for claims adjusters and underwriters.

⚡ Key Takeaways

  • Manual claim review exceeds 100 minutes per case (baseline).
  • Xtrakto.AI performs document ingestion, splitting, classification, extraction, enrichment, and post‑processing without pre‑configured templates.
  • EXL Insurance LLM is fine‑tuned on insurance and medical data to understand clinical terminology, ICD/CPT codes, and diagnosis‑treatment relationships.
  • The solution runs on Amazon SageMaker AI, leveraging multi‑GPU fine‑tuning and managed inference.
  • No separate template engineering is required; the pipeline is template‑agnostic.
💡 Why It Matters

Engineers can replace costly, error‑prone manual review pipelines with a unified, AWS‑native IDP + LLM stack, cutting processing time and improving consistency for high‑volume, regulated medical document workflows.

✅ Practical Steps

  1. Deploy Xtrakto.AI on AWS to handle document ingestion, splitting, classification, extraction, enrichment, and post‑processing.
  2. Fine‑tune and serve the EXL Insurance LLM using Amazon SageMaker AI’s multi‑GPU training and managed inference capabilities.
  3. Integrate the two components into a single pipeline that feeds extracted data from Xtrakto.AI into the LLM for summarization, querying, and structured output generation.

Want the full story? Read the original article.

Read on AWS ML Blog ↗

More like this

Censored vs. Truncated Regression in Python, Explained

Towards Data Science•#llm

New in AWS Quick: More capable flows, an updated web experience, and agents that work together

AWS ML Blog•#agents

Three insights you may have missed from theCUBE’s coverage of the Dell Technologies New York Symposium

SiliconANGLE AI•#deployment

How Google Duplex evolved to also call businesses for you in Gemini

Google Research•#agents

EXPLORE AI NEWS

Daily hand-picked stories on LLMs, RAG, agents and production AI — curated for engineers who ship.

BROWSE NEWS

GET THE WEEKLY DIGEST

Join engineers getting the Monday signal-over-noise AI breakdown. No spam, unsubscribe anytime.

LEARN AI ENGINEERING

Curated courses, research papers, repos and tutorials built for engineers leveling up in AI.

START LEARNING