AI Intake Chatbot with Multi-Language Support for a Community Health Center

Overview

Built a multi-language AI intake chatbot for a community health center serving a patient population where a significant share of patients are more comfortable in a language other than English. The chatbot conducts pre-visit intake in the patient's preferred language and translates the structured summary into English for clinical staff, reducing reliance on scheduling live interpreters for routine intake questions. It was built with a conservative scope, focused on structured intake information rather than open-ended medical conversation, and always escalates to staff for anything ambiguous. The center's front-desk team was involved throughout the build to make sure the tone and question set matched how they actually talk to patients. A community health center serving a large non-English-speaking patient population relied heavily on scheduling live interpreters just for routine pre-visit intake questions, tying up interpreter time needed elsewhere. We built a multi-language AI intake chatbot, scoped conservatively to structured intake rather than open-ended medical conversation, that conducts pre-visit intake in the patient's preferred language and translates a clean summary into English for clinical staff. We worked closely with the center's front-desk team to make sure the chatbot's tone and question set matched how they actually talk to patients, not a generic script. After testing translation accuracy across the center's most common languages, we piloted it with one language before expanding to the full set. Routine pre-visit intake no longer requires scheduling a live interpreter, freeing that time for visits where it's genuinely needed while keeping non-English-speaking patients well-served at intake.