Matt Mayo

I'm building Cashe — an AI agent — plus a small fleet of specialist agents behind it, hands-on and from the ground up. This site is the public record of that work: what I'm building, why, what I think it means, and the receipts along the way.

What I'm doing

Designing, running, and observing autonomous AI agents in the open. The centerpiece is Cashe — an agent I'm developing incrementally as a learning vehicle — surrounded by a fleet of task-specific agents (a supply-chain paper-trading desk, a crypto desk, a research agent, a slide-deck builder, a watchdog, and more). I'm not scripting them rigidly; I'm watching how they actually behave, and iterating.

Why I'm doing it

To build real, working expertise in AI and autonomous agents — not by reading about them, but by running them. The fastest way to genuinely understand what these systems can and can't do is to stand them up, give them problems, and study the results. The Cashe Log is that experiment's notebook, published as-is — the receipts, not a highlight reel.

By day, I'm Chief Risk Officer at Community Bank, where I've spent nearly three decades across lending, banking operations, and now enterprise risk — including watching AI move from hype to operational necessity inside a real bank (loan review, check-fraud detection, and the vendor systems that quietly embed it whether you ask for it or not). That vantage point shapes a lot of how I think about building agents here. More on that on the About page.

What I believe about AI

• [DRAFT — needs Matt's actual words] AI isn't a separate topic from work; it's about to be folded into how nearly everyone does nearly everything.
• [DRAFT — needs Matt's actual words] The people who'll do best with this shift are the ones who get hands-on early and build judgment, not the ones who wait for it to be "figured out."
• [DRAFT — needs Matt's actual words] Agents are the next step beyond assistants — systems that act, not just answer — and the gap between "demo" and "dependable" is where all the real work is.

Explore