The challenge
Bramwell Health's patient support inbox received a high volume of repetitive questions about appointment scheduling, billing, and insurance coverage — questions that had clear, documented answers but still consumed most of the support team's day, delaying more urgent patient issues.
What we built
We built a retrieval-augmented generation (RAG) assistant on top of Anthropic's Claude, grounded in Bramwell's own policy documents, FAQ content, and scheduling system via a secure API integration. The assistant answers directly where confidence is high and hands off to a human agent with full context where it isn't — with every automated response logged for review.
The results
Within the first quarter live, the assistant resolved roughly a third of incoming tickets without human involvement, and average response time for patients dropped 40% as the support team's queue shrank. The system was scoped, built, and in production in under eight weeks, with a defined escalation path that kept clinical staff comfortable with what the assistant was and wasn't allowed to answer.
Simbyte