Case Study: Major US Healthtech achieves explainable, deterministic oncology treatment recommendations with Ideas2IT Technologies

Major US Healthtech builds 8-step oncology treatment pathways with Ideas2IT Technologies

Ideas2IT Technologies built an Agentic AI system for a major US healthtech company, a large public health system in South Florida. The challenge was to help oncologists efficiently determine treatment regimens from complex guideline PDFs, which contained intricate branching logic and clinical qualifiers. The lack of a system to handle inconsistent biomarker data and prior treatment history across multiple lines of therapy created significant decision friction at the point of care.

The solution implemented by Ideas2IT Technologies was a two-pipeline AI system on Azure AI Foundry. It first converted unstructured guideline PDFs into structured decision trees in a batch process. A real-time inference pipeline then used deterministic entity matching and an 8-step agentic sequence to produce explainable treatment recommendations. The results included a reliable platform that delivers auditable, constraint-enforced regimen shortlists with explicit reasoning for both selections and rejections, effectively traversing multi-line treatment scenarios for oncologists.