AXA Indonesia

AXA Indonesia

Transforming AXA's Claims Process: Reducing Errors & Accelerating Decision Making

Transforming AXA's Claims Process: Reducing Errors & Accelerating Decision Making

Transforming AXA's Claims Process: Reducing Errors & Accelerating Decision Making

90% reduction

Errors made by employees

45% faster

Claims processing

75% accuracy

Of the AI decision suggestions

My Role

Product Designer

The Team

1 x Product Designer

1 x Product Manager

1 x Head of AI Engineering

2 x AI Technology Strategist

2 x AI Engineers

1 x Delivery Manager

Year

2024

Problem

Problem

Problem

Imagine managing claims for over 700,000 insurance customers while relying on manual processes prone to delays, inconsistencies, and errors. This was the reality for AXA Indonesia, a leading branch of the global AXA Group offering medical insurance solutions.


Their post-authorization claim review process demanded extensive manual effort from 50+ medical officers who analyzed detailed PDFs containing treatment procedures, medications, and expenses. These officers assessed the claims against international medical guidelines to determine their legitimacy.


Key challenges we've observed were:


⌚ Time-consuming reviews

Lengthy approval timelines caused frustration among customers.


🤔 Inconsistent Decisions

Varied interpretations of medical guidelines led to potential errors.


Fraudulent Claims

Significant costs incurred from overtreatment and fraudulent claims.


AXA needed a transformative solution to address these inefficiencies and improve their claim review process.

Imagine managing claims for over 700,000 insurance customers while relying on manual processes prone to delays, inconsistencies, and errors. This was the reality for AXA Indonesia, a leading branch of the global AXA Group offering medical insurance solutions.


Their post-authorization claim review process demanded extensive manual effort from 50+ medical officers who analyzed detailed PDFs containing treatment procedures, medications, and expenses. These officers assessed the claims against international medical guidelines to determine their legitimacy.


Key challenges we've observed were:


⌚ Time-consuming reviews

Lengthy approval timelines caused frustration among customers.


🤔 Inconsistent Decisions

Varied interpretations of medical guidelines led to potential errors.


Fraudulent Claims

Significant costs incurred from overtreatment and fraudulent claims.


AXA needed a transformative solution to address these inefficiencies and improve their claim review process.

The solution

The solution

The solution

AI-Powered Claim Analysis Tool

AI-Powered Claim Analysis Tool

AI-Powered Claim Analysis Tool

Our team stepped in to design and deliver a proprietary solution powered by Generative AI and Large Language Models. The tool combined cutting-edge AI technologies with robust datasets from Wolters Kluwer UpToDate and DrugBank, tailored specifically for AXA’s needs.


After a series of user interviews, analysis on current state of operations and AXA's competitors, we've developed the strategy to transform their claims processing with:


🤖 Automated Claim Analysis

The AI efficiently processed claims, applying medical guidelines consistently.


🧠 Human-in-the-Loop (HITL) Mechanism

Integrated human oversight ensured continuous learning and improved accuracy.


📈 Knowledge Graphs

Built-in clinical guidelines enhanced reliability, reducing errors and fraudulent claims.


🗃️ Data-Driven Insights

Comprehensive analytics identified trends in claims data for strategic decisions.

Impact

By combining advanced AI automation with human expertise, the solution radically improved AXA Indonesia’s operations.

By combining advanced AI automation with human expertise, the solution radically improved AXA Indonesia’s operations.

45% faster

Claims processing

45% faster

Claims processing

45% faster

Claims processing

90% reduction

Errors made by employees

90% reduction

Errors made by employees

90% reduction

Errors made by employees

75% accuracy

Of the AI decision suggestions

75% accuracy

Of the AI decision suggestions

75% accuracy

Of the AI decision suggestions

Cost savings

Reduced payouts for unnecessary or fraudulent treatments

Cost savings

Reduced payouts for unnecessary or fraudulent treatments

Cost savings

Reduced payouts for unnecessary or fraudulent treatments