Proof

I’ve been building business systems since long before anyone called them AI.

Two flagship cases and three smaller ones. All of them are my own ventures and my own work — which means I can show you how they were actually done.

01 / The operating company

Distillery 291

Eleven years as COO/CFO · internationally recognized craft whiskey producer

The situation: a founder-led craft producer with the classic pattern — data sprawled across systems that didn’t agree, workflows held together by individual effort, and software subscriptions that fit 80% of the need while the missing 20% created most of the friction.

What I did: ran finance and operations — capital, people, pricing, multi-state distribution, facilities — and built the operating systems along the way, each at the smallest tool that solved the problem:

  • A custom CRM, evolved across three platforms as the company’s needs changed
  • Inventory systems integrated with point-of-sale
  • A shipping request-and-tracking tool that replaced a paid subscription outright
  • Commission calculators, cash-management apps, and a daily reporting rhythm run with a remote contractor I hired and managed

The result: a company operable beyond the founder’s personal bandwidth — and when I transitioned out, every system went with a written runbook, so nothing depended on my memory.

02 / The AI platform

CouncilWatch (CivixIQ)

Founder · designed, built, and operated — with paying subscribers

The situation: government meetings are public but practically opaque — hours of video, dense agendas, votes nobody tracks. I set out to make one person able to do the work of an analysis team, at a quality level people would pay for.

What I built: an AI system that turns the raw meeting record into published analysis and scoring. Roughly half the pipeline is careful data engineering; the other half is structured AI reasoning — with the checks, second looks on high-stakes items, and testing-before-changes that make the output trustworthy enough to sell. When the AI is wrong, there’s a process to find it, correct it, and log it.

Why it matters to you: every reliability problem your company will hit adopting AI — where to trust its judgment, how to catch silent errors, how to change things without breaking what works — I’ve already hit in production, on my own money.

The result: 100+ meetings analyzed across two governments, a multi-year scored dataset, paying subscribers — run by one operator.

CouncilWatch vote detail page: an AI-drafted summary of a city council vote, the outcome and tally, and the case for voting yes and for voting no
03 / The smaller stories

Three that show the range.

Different problems, same pattern: a material outcome, honest about limits, willing to do the unglamorous work.

The $200K nobody wanted to touch.

Outside tax professionals declined Employee Retention Credit work — too much bad practice in that market. I did the work myself, documented exactly what I was and wasn’t qualified to attest, and secured over $200,000 in credits for the company.

The subscription we stopped paying for.

We were paying for a project-management platform mostly to run one shipping workflow. I rebuilt it as a purpose-built tool — request form, coordinator queue, live tracking — and canceled the subscription. Immediate, measurable ROI.

Financing the barrel, not the bottle.

Whiskey ages two years before it earns a dollar. I designed an inventory-collateralized financing structure so the aging stock itself carried the growth.

Also on the record: 15 years as a Wall Street Journal–ranked equity analyst · CFO of a Denver entrepreneurship ecosystem · President of the Denver chapter of Keiretsu Forum, the world’s largest angel investor network · mentor and advisor to early-stage companies, from organic wine (Wander + Ivy) to enterprise analytics (PieAX).

Your operation has stories like these waiting.

The assessment finds them — where the value is, what it costs you today, and what it takes to capture it. It starts with a conversation.

Book a 30-minute conversation Free. Confidential. No commitment required.