Rachel is a voice AI agent developed at the ETH Student Project House that answers inbound support calls in German and Swiss German — resolving routine inquiries autonomously and handing complex cases to humans with full context.

We have built a working voice AI prototype and validated the problem in over a dozen discovery conversations with Swiss companies, supported by structured B2B outreach and a fully Swiss-hosted architecture as core differentiator. Our main goal now is to scale on the client side: converting conversations into pilot projects and building our first reference customers.

The project started with Peter, who was already building a voice AI prototype on his own. He and Philip reconnected via email — they knew each other from previous ETH projects — and Philip was immediately drawn to the topic, having heard about exactly these challenges from customers in his B2B sales work. With Philip's sales background and Peter's machine learning expertise, one piece was still missing: Adrian, who was already running his own IT consulting business with clients.