ZEISS Medical Technology

Solution

BCG worked with ZEISS Medical Technology to build a proof-of-concept generative-AI application for ophthalmology practices. After assessing a hundred possible GenAI use cases, the joint team selected patient-inquiry response generation and built the first version in eight weeks. The application ingests patient-submitted content forms and self-test forms, retrieves relevant facts from a library of approved Zeiss marketing and patient materials, and uses Azure OpenAI with LangChain, vector databases, and LLM agent techniques to draft responses. BCG and the Zeiss AI team refined the workflow in agile iterations using customer insights and tester feedback, while responsible-AI guardrails and human validation checked output quality. The result is a validated assistant that can produce personalized, patient-ready answers while staying grounded in approved source content.

Data flow

Patient-submitted content forms and self-test forms enter the application, which retrieves relevant passages from approved Zeiss marketing and patient materials stored for RAG. LangChain and LLM agent logic orchestrate the workflow, Azure OpenAI generates draft responses, and validated outputs are returned to practices as patient-ready answers or drafts for review.