Digitizing Handwritten Pharmacy Records with AI | High-Accuracy Medical VLM Case Study

Handwritten pharmacy records are still a bottleneck in healthcare operations — slow, error-prone, and hard to scale. This session shows how AI actually solves it in production, not in theory. ⏱️ Key moments: 00:00 Why handwritten pharmacy scripts are still a problem 02:40 Manual workflows and operational bottlenecks 06:10 AI pipeline: from scanned images to structured clinical data 11:20 How the hybrid VLM + NLP system works in practice 17:30 Accuracy, compliance, and real-world deployment challenges Nima Babazadeh (Enterprise Data & AI Architect, EBOS Group) and Jiri Dobes (Head of Solutions, John Snow Labs) walk through how millions of handwritten pharmacy scripts are digitised using a hybrid medical VLM pipeline. At EBOS Group, one of the largest healthcare distributors across Australia, New Zealand, and Southeast Asia, the challenge was not experimentation — it was accuracy, compliance, and operational reliability at scale. Generic LLMs failed to meet requirements. What worked instead was a deterministic, multimodal architecture combining Visual NLP, medical LLMs, and healthcare-specific NLP. The result: a production-grade system that extracts, validates, and structures handwritten prescriptions in minutes while meeting strict data residency, auditability, and clinical safety requirements. 📌 Applied Healthcare AI Summit 2026 — real-world AI deployment in healthcare, from pilots to production systems. #HealthcareAI #MedicalAI #AIinHealthcare #OCR #HandwritingRecognition #MultimodalAI #VisualLanguageModels #ClinicalData #PharmacyTech #AIGovernance #DataPrivacy #RealWorldAI #HealthcareOperation