Digitising your claims data: Unlocking the future of health claims automation
- Manual claims processing remains a significant challenge for Australian private health insurers despite high automation in point-of-service claims, impacting member experience and operational efficiency.
- Digitising member-submitted claims data using Optical Character Recognition (OCR) is the crucial first step towards unlocking claims automation.
- Modern OCR, powered by Generative AI, overcomes the limitations of earlier technologies by offering faster, more accurate data extraction without the need for extensive training on sensitive member data.
- Accurate data extraction paves the way for automating more complex aspects of the claims process, ultimately leading to improved member satisfaction, reduced costs, and future innovation.
Current landscape
The claims process is critically important for private health insurers. A smooth, efficient and transparent claims process is crucial for member satisfaction and retention, operational efficiency, cost management, reputation and brand image, regulatory compliance, personalisation and maintained viability.
Claims automation leverages technology to process insurance claims with minimal human intervention.
While major Australian health insurers have achieved high automation rates for point-of-service claims, as high as 96% across 20 million successful claims per year, manually submitted member claims continue to demand significant human effort, representing 1 million claims annually. This is where we see particular challenges arise around pharmacy invoices with inconsistent formats, hand-written invoices from specialists and potentially nefarious circumstances where members upload photos of anything to get the claim submitted.
The current landscape of manual claim submissions varies across insurers. Some prioritise customer experience by requesting minimal initial information, while others place the burden of data entry on the member, and in return offering limited transparency and communication throughout the claims process. The implications of a better member experience upfront often leads to longer processing times later, while heavy upfront data entry can lead to member frustration and potentially churn.
The typical member experience:
Since 2016, Mantel has partnered with Australia’s leading Private Health Insurers (PHIs), empowering them to transform from traditional funders into innovative payviders. Our experience tells us that positive claims experiences for members builds trust and loyalty, leading to higher member retention rates.
For member submitted claims, a frictionless digital experience for members involves more proactive preparation pre-claim, a more seamless experience for claims initiation (i.e. simple photo upload of their invoice via the member mobile app or web portal) and leaving the system to handle the rest. This not only minimises the risk of data entry errors but also significantly enhances customer satisfaction.
A faster and more personalised claims experience:
The power of intelligent data extraction
Achieving the ultimate claims process for members is not a small undertaking. Based on our experience working with other PHI’s in this space we recommend starting with digitising claims data at the point of member submission.
Realising this seamless experience requires a robust Optical Character Recognition (OCR) solution capable of accurately digitising claims data and seamlessly feeding it into the claims workflow.
Early forays into applying OCR to the claims process around 2022 has left insurers sceptical. These first-generation technologies demanded the costly and time-consuming creation of custom machine learning models, requiring vast amounts of labelled data – raising significant Personally Identifiable Information (PII) concerns during the training process due to the need to expose sensitive health information for model training.
Today, the landscape has dramatically shifted. The latest advancements in OCR, combined with the power of Generative AI are delivering far more promising results with its ability to understand natural language and visual context. By leveraging GenAI alongside traditional extraction techniques, businesses can now implement faster, more accurate, and more efficient data extraction methods. These advanced vision models possess a deep understanding of context, significantly reducing or even eliminating the need for extensive upfront training with sensitive customer data, bolstering privacy and security.
Example results
We submitted the following documents; with differing levels of difficulty, to our Intelligent Docs Automation solution to demonstrate the accuracy we’re able to achieve.
Easy: PDF document
Fields correctly extracted
Issuer name: Yes
Invoice amount: Yes
Invoice date: Yes
Invoice number: Yes
Total ex GST: Yes
GST amount: Yes
Line items: Yes
Accuracy rate
100%
Speed of extraction
Less than 7 seconds
Medium: Photo of receipt
Fields correctly extracted
Issuer name: Yes
Invoice amount: Yes
Invoice date: Yes
Invoice number: Yes
Total ex GST: Yes
GST amount: Yes
Line items: Yes
Accuracy rate
100%
Speed of extraction
Less than 7 seconds
(Yes, this is the original quality)
Hard: Photo of handwritten document
Fields correctly extracted
Issuer name: Yes
Invoice amount: Yes
Invoice date: No (2027, instead of 2022)
Invoice number: Yes
Total ex GST: Yes
GST amount: Yes
Line items: Yes
Accuracy rate
85.7%
Speed of extraction
Less than 7 seconds
Bonus: Photo of a dog.
Sometimes members accidentally upload the wrong image.
Fields correctly extracted
Issuer name: N/A
Invoice amount: N/A
Invoice date: N/A
Invoice number: N/A
Total ex GST: N/A
GST amount: N/A
Line items: N/A
Accuracy rate
The system correctly identified that this is not an invoice, and an error would be presented to the member.
Speed of extraction
Less than 7 seconds
Beyond extraction: what’s next?
With confident and accurate data extraction as the foundation, the next frontier is automating the more complex aspects of the claims process, including identifying & retrieving missing information, fraud detection and automated payments.
Ready to unlock the future of Claims Automation?
By taking the crucial first step of digitising your claims data with Mantel’s intelligent data extraction solutions, you can lay the groundwork for a streamlined, efficient, and member-centric future. Contact us today for a consultation and discover how we can help you unlock the full potential of claims automation.
