We research and build AI systems for finance and payments. Our work spans payment automation, agent infrastructure, and the thorny problems that emerge when language models handle real money.

AI systems that understand payment workflows end-to-end. We're exploring how language models can handle reconciliation, fraud patterns, and exception handling.
In development with design partners
Tools for connecting AI agents to existing systems. We build the plumbing that lets language models work with real APIs and real data safely.
Production systems on Google Cloud
Research into how AI systems represent financial brands and make subjective assessments. We study bias patterns and build evaluation frameworks for high-stakes domains.
Active research area
The infrastructure for AI in finance is being built right now. Most of the hard problems aren't about the models themselves—they're about trust, integration, and making systems work reliably when real money is involved. That's where we focus.