Continuum AI Concierge

I developed, tested, and rolled out an AI concierge for a client using Python, Retell AI, and FastAPI to manage a steady flow of customer phone calls tied to a building’s parking renovation project. The goal was to create a smooth, always-available first point of contact that could confidently handle the repetitive “day-to-day” questions people tend to ask during construction—things like where to park, what areas are temporarily closed, expected timelines, and what to do in common situations—while keeping the experience friendly and consistent.

On the technical side, I built the backend service with FastAPI, integrated it with Retell AI for real-time voice interactions, and structured the conversation logic so the assistant could collect key details, route the caller through the right path, and provide accurate, easy-to-understand information without bouncing people around. I also put emphasis on reliability and maintainability: clear endpoints, predictable workflows, and a setup that makes it straightforward to update messaging as renovation phases change.

As a result, the concierge successfully automated a large portion of routine inquiries, reduced back-and-forth for the building team, and helped free up staff time for more complex, high-touch requests—ultimately lowering overall workload by about 20% while improving the speed and consistency of customer responses.