Myriad Genetics and AWS: Revolutionary AI for Diagnostics

Myriad Genetics and AWS: A Collaboration to Optimize AI

Every year, Myriad Genetics, a company specializing in diagnostic testing, manages approximately 1.5 million tests. These tests include hereditary cancer screening, prenatal conditions, and tumor profiling. The Salt Lake City-based company has relied on artificial intelligence to accelerate the processing of documents necessary for billing, a complex process known as revenue cycle management. However, the initially developed tool turned out to be costly.

To address these challenges, Myriad collaborated with Amazon Web Services (AWS) to rebuild its AI tool, named Image Genius, using Amazon Bedrock. This partnership allowed some Myriad employees to save up to 300 hours per month, significantly improving the company's efficiency.

Myriad Genetics also handles assessments of how genes can affect the body's response to mental health medications. It can take the company months to process all the paperwork necessary for billing, a process that requires healthcare organizations like Myriad to dedicate staff to software, medical coding, and regulatory compliance to manage the billing of medical services.

An Urgent Need for Technological Transformation

In its effort to optimize operations, Myriad Genetics developed an internal AI tool aimed at automating diagnostic, insurance, and reimbursement processes. However, the first version of this tool, launched in 2024, proved too costly to maintain. Faced with this situation, Myriad sought AWS's help to reduce the costs associated with this ambitious project.

With the integration of new AI capabilities, document processing time was dramatically reduced from 10 minutes to just 20 seconds per document. This improvement accelerated reimbursement cycles, a crucial aspect for the financial health of the company. Martyna Shallenberg, Senior Director of Software Engineering at Myriad, emphasized the importance of receiving payments more quickly, a process that can take up to 18 months due to the complexities of insurance claims.

The Technology Behind the Transformation

In the fourth quarter of 2023, Myriad conceptualized a tool capable of using AI to classify pathology reports, automate document tagging, and extract keywords from Medicare. Kevin Haas, former CTO of Myriad, played a key role in this initiative before leaving the company to join Veracyte.

The goal was to expedite the prior authorization process, where a physician must obtain approval from an insurer before ordering a test for a patient. This process is crucial as a quick prior authorization is often necessary to ensure health insurance coverage. Without this authorization, claims could be denied, delaying patient care and impacting Myriad's revenues.

In the second quarter of 2024, Haas and Myriad worked with consulting giant PwC to build the first version of its intelligent document processing tool, called Image Genius, which utilized the Amazon Textract machine learning service and the Amazon Comprehend natural language processing service to automate document classification, extract data from pathology reports, process insurance cards, and reduce manual document handling. Myriad deployed Image Genius for its hereditary cancer business in the third quarter of 2024 to automate the classification of medical documents and streamline revenue cycle workflows.

Challenges and Solutions: A Strategic Pivot

In 2024, as Myriad planned to expand Image Genius to other business units, it encountered issues related to costs, longer-than-expected processing times, and difficulties in extracting the right clinical information from dense pathology reports. Haas indicated that Myriad had determined that the internally developed AI tool required constant maintenance and updates. "The inputs are as vast as the millions of samples we receive each year," Haas stated.

Myriad discovered that the AI system was not always clever enough to differentiate a test order from a physician's medical notes. "That was actually the hardest part of the process, separating those documents," Shallenberg added. Myriad spent the first quarter of 2025 refining the AI model prompts to improve the accuracy of its results. They also provided more data samples to give the system greater context but decided that the results of this current iteration of Image Genius were not worth the cost. AWS Comprehend, which uses machine learning to identify clinical notes, pathology reports, insurance information, and other data from unstructured text and data, was very expensive.

A Redesign with AWS for a Promising Future

In the spring of 2025, Myriad pivoted by collaborating with the AWS Generative AI Innovation Center to rebuild Image Genius on Amazon Bedrock. This new architecture allowed Myriad to test different AI models and choose the one that offered the best balance between speed, accuracy, and cost.

Dr. Rowland Illing from AWS emphasized the importance of selecting the right model to optimize performance. This redesign resulted in a next-generation platform, finalized in the third quarter of 2025 and relaunched in early 2026.

Tangible Results and Future Prospects

Thanks to this redesign, Image Genius has enabled Myriad to save approximately 300 hours per month in the women's health division, processing around 9,000 prior authorizations. The tool will be gradually deployed in the oncology and mental health divisions by 2027.

Under the leadership of new CTO Raj Jampa, Myriad aims to further integrate AI and automation into all aspects of the revenue cycle, including eligibility verification and claims management. This strategy promises to transform Myriad's operations, strengthening its position in the medical diagnostics sector.