AI/ML Cures Manual Processing Pains of Donor Prescription Program

Heart to Heart International (HHI) equips teams, organizations and clinics with the supplies needed to make a real difference. Through the Custom Order Medical Aid program they help more than 250 nonprofit partner organizations select specific medicines and medical supplies for short-term or long-term domestic and international medical projects. Since 1992 HHI has distributed more than $2.4 billion in humanitarian aid worldwide by connecting available aid from strategic donor partnerships with leading manufacturers like Abbott, AbbVie, BD, Henry Schein, Johnson & Johnson, Mylan, Pfizer and Sanofi.

Since the custom medical aid program provides prescription drugs, the request process must be thoroughly documented and can require multiple types of forms to be filled out by the patient to evaluate their candidacy for prescription support. The manual process included more than 200,000 pages of data per month, which were hand-coded into HHI’s CRM application and then routed for evaluation and approval. Due to the manual process, human error impacted processing time and surge periods for prescription renewals required additional seasonal help. Those costs diminished the impact of donations.

Download the Case Study

Technologies Involved


Intelligent Document Processing


Artificial Intelligence/Machine Learning
AWS Textract
S3 Encryption
AWS Lambda Services
AWS CloudWatch
AWS Simple Email Services


Key Pain Points

Fax submission of paper forms by patients created resource equity roadblock.

Salesforce contact lookup and manual input of values opened door to inaccurate interpretation of unclear form data, and
errors would block prescription fulfillment.

Volume of documents – more than 200,000 pages per month – drained resources and diverted charitable impact.

The AWS team at TekStream deployed AI/ML services to automate capture and validate form data using a customized Intelligent Document Processing engine. Forms are aligned and OCR operations on text, form data, table information, signature detection and keyvaluepair data extraction are performed. The full form package and JSON of all relative data are sent to Salesforce, where a custom interface highlights indicators of questionable data using geometry location tags and artificial intelligence confidence on the accuracy of collected values. The coder can now quickly scan raised issues and push to approval in seconds instead of hours.

In early 2024, Heart to Heart International shared that the TekStream Intelligent Document Processing project helped them attain all 5 of their KPI’s for the first time in three years.

Key Successes

  • Time to process reduced from hours to seconds — increasing their ability to service patients in need faster and more accurately.
  • Data geometry and AI driven confidence scores output allows machine processing of errors, encrypted inbound, during storage and outbound.
  • Accuracy incentivizes drug manufacturer partners to continue or increase contributions due to improved humanitarian aid impact.
  • Cost-savings reduce organization overhead and increase nonprofit operating margin ratio.

Customer: Heart to Heart International (RegaloRX)

Industry: Healthcare Nonprofit Organization

HQ: Lenexa, Kansas

Heart to Heart International (HHI) equips teams, organizations and clinics with the supplies needed to make a real difference.