Hey! Saner, Yash, and Kayla from Trace here.

TL;DR: Trace (tracetec.co) builds voice AI agents to handle customer support for fintechs and banks. Beyond answering generic questions, our AI agents can actually take actions like "replace my card" or "dispute this charge" while maintaining human-like conversation.

Here is a quick demo

Problem

Customer service is one of the largest operational spends across banks (15% opex), insurance (40% opex), and fintechs. However, call centers offer inconsistent service quality despite the cost.

Existing AI solutions only read off of public documents and deflect the user — they can’t access account-specific data or take actions like human agents would. In financial services, this causes extreme user frustration.

Solution

Trace AI agents securely integrate with core systems to offer instant, 24/7 resolution over the phone. So far, we achieve 95% end-to-end resolution of calls across quoting, card applications, transaction tracking, and more — all while following policy and never leaking sensitive data. With under 0.2s time to first token, we are also optimized for seamless interactions over voice.

Team

We are all ex Stanford/Berkeley researchers, specialized in building reliable AI solutions for high-stakes applications. Saner (CEO) built an AI copilot that helped UCSF cardiothoracic surgeons save real lives. Yash (CTO) was a quant researcher at Optiver and Goldman Sachs, where he enabled millions in annual profitability with his high frequency trading strategies. Kayla (COO) led a project at the Berkeley NLP Group that laid the foundation for Trace’s tech.

Ask

If you know any decision-makers at:

• fintechs
• banks
• insurance providers
• BPOs,

please contact founders@tracetec.co — we’d love to get their perspective!