Skip to main content




Thea – AI Document Processing
Machine Learning
Artificial Intelligence
Data Science

Key Takeaways



Savings in manual handling costs over 12 months


Accuracy for key fields from medical receipts


Touchless claims


Further reduction in manual handling time on non-automated claims

Customer Overview

nib is a health and medical insurance provider in Australia with approximately 1,500 employees and over 1.6 million members

Business Challenge

The nib mobile app allows members to easily make claims by submitting a photo of a receipt from their mobile device. Relevant information from submitted receipt photos was manually entered into the nib claims system.

Significant time was spent on manual data entry on top of assessing the claim. As the app gained popularity nib realised this approach was not scalable.

Success Criteria

  • Integrate an automated solution to assist nib employees with the manual entry of key fields from the receipts
  • Seamlessly fit into nib’s current business processes and technology
  • Sensitive receipt data must remain within nib’s secure AWS environment at all times
  • Reduce the turnaround time for each member by enabling the claims team to process assessments faster


Amazon Textract was selected for its accuracy and ability to determine the document structure while extracting text. Geometric de-skewing and key-value pair recognition were particularly effective in handling claims receipts.

Mantel Group created a image processing pipeline to:

  • Find and extract relevant fields from receipt photographs
  • Analyse the extracted text and identify information required for the claims process
  • Pre-populate the claims processing system, ready for employee review
  • Provide nib with automated monitoring to validate accuracy of the solution