Why a COO hires me instead of Deloitte or a junior in-house AI lead.

The receipt: what I shipped at Paytronix.

I spent five-plus years at Paytronix Systems, a restaurant loyalty and payments platform. My last role there was Director of Data Insights Productization — leading an eight-person team building the data systems that powered analytics for Paytronix’s clients across the restaurant industry.

The work was applied. We weren’t building models in isolation; we were building reporting pipelines that real operations teams relied on to make decisions in real time. One of the things that sticks with me most is a reporting environment rebuild where we took a process the team was spending 14 analyst-hours a week on and got it to three. That’s 11 hours a week back — redirected toward work that actually required human judgment.

Before Paytronix, I was at The Economics Center, which is where I learned that data work only matters if it changes what someone does. A technically correct analysis that nobody acts on is just an expensive document.

I Run What I Sell

The prospecting system I’d build for your sales team is the same one I’m running for my own practice right now. Same architecture, same targeting logic, same enrichment pipeline. I built it for myself first, ran it, found what breaks, and fixed it. That’s the version I’d build for you.

I’ve also been building AI agent infrastructure for my own use — a working system with real components: memory management, automated decision routing, governance and safety controls, a build pipeline. It’s not a demo. I run it daily, I know where it breaks, and when I scope AI systems work for a client, that implementation experience is what I’m drawing on.

If I’m recommending it, I’ve run it.

How the system actually runs →

One Operator. No Handoffs.

I work alone. The person who scopes your project, prices it, builds it, and hands it off is me — not a senior consultant who sells the engagement and then disappears while a junior analyst does the work.

Working with a one-person practice means real continuity. I know what we decided in week one when we’re in week six. I know why the pipeline is built the way it is, because I built it. When something unexpected comes up — and something always comes up — you’re talking to the person who can actually resolve it.

The tradeoff: I work with a small number of clients at a time. If your timeline is “we need a 20-person team by next Monday,” I’m not your answer.

Field notes

Working notes on agent systems, governance, and what happens when you actually run AI in production — at the development site. The Dolphin Agent series is the most honest representation of how I work through hard problems.