Hands-free first notice of loss: Using Strands Agents and Amazon Bedrock AgentCore Browser Tool for intelligent claims intake

2026-06-09 · Source: Artificial Intelligence · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Cloud Computing & IT Infrastructure · Depth: Intermediate, extended

Summary

A hands-free first notice of loss (FNOL) intake system combines Strands Agents SDK with Amazon Bedrock AgentCore Browser Tool and Amazon Nova Act to automate the processing of multimodal evidence in insurance claims. This solution interprets, validates, and correlates unstructured data like photos, videos, documents, and audio notes, transforming them into tagged, decision-ready inputs. Strands Agents perform domain-specific reasoning using foundation models via Amazon Bedrock, while Nova Act orchestrates browser interactions through a managed Chrome session provided by AgentCore Browser Tool. The system assesses claim complexity, categorizing claims as "Simple" or "Complex," and provides adjusters with context-rich, pre-analyzed submissions, aiming to reduce manual screen work, improve claim cycle times, and enhance customer experience without modifying existing portals.

Key takeaway

For Claims Operations Managers seeking to reduce manual intake validation and accelerate claim processing, implementing an agentic FNOL system like the one described can significantly improve efficiency. You can utilize existing portals without modification, allowing adjusters to focus on judgment rather than repetitive screen work. This approach provides structured, tagged evidence for faster routing and more consistent decision-making, especially during volume spikes.

Key insights

Automated FNOL intake combines browser automation with domain-specific AI agents to structure multimodal evidence and accelerate claims processing.

Principles

• Separate UI orchestration from domain reasoning.
• Encode domain knowledge into agentic business rules.
• Generate audit trails as a byproduct of execution.

Method

Nova Act drives portal interaction via AgentCore Browser Tool, while Strands Agents apply FMs on Amazon Bedrock for evidence interpretation, cross-modal correlation, and claim complexity assessment.

In practice

• Automate FNOL intake for insurance claims.
• Tag multimodal evidence at ingestion.
• Triage claims based on complexity.

Topics

• First Notice of Loss
• Generative AI Agents
• Amazon Bedrock AgentCore
• Amazon Nova Act
• Insurance Claims Automation
• Multimodal Evidence Tagging

Code references

• aws-samples/sample-browser-automation-with-agentcore-for-insurance-fnol-claims-queue

Best for: AI Engineer, MLOps Engineer, AI Architect

Related on AIssential

• Agents that transact: Introducing Amazon Bedrock AgentCore payments, built with Coinbase and Stripe — AI agents can now autonomously transact for resources using integrated, managed payment capabilities.
• Introducing OS Level Actions in Amazon Bedrock AgentCore Browser — OS Level Actions in AgentCore Browser enable AI agents to interact with native OS UI elements.
• AWS, Coinbase & Stripe Launch Autonomous Paying AI Agents — AWS Bedrock AgentCore Payments enables AI agents to autonomously transact for digital resources using Coinbase and Stripe.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence.