I drive AI adoption. And I ship the thing.

Why this profile is rare

Three things that almost never come together

Organizations hiring for AI transformation usually get one of these. Adoption stalls because the advisor cannot build, or the builder cannot bring people along. I do both, and I evaluate the tools with a discipline most vendors would rather you skipped.

Built and live

Three products I designed and built

Each of these went from idea to live product with no dev team behind it, and each one is open on the web right now: click through and judge for yourself. An honest status note: all three are live and deployed, and none has been through a formal beta with real users yet. That work is next, and I would rather say so than dress it up.

Beyond the products

The operation behind them runs itself

The three products above are the visible output. The harder build, and the one a business leader should care about, is the operation underneath: a one-person company run with the checks and routines you would normally need a team for. In plain terms:

• The overnight shift runs unattended. Every night, across multiple machines: repositories synced and backed up, automated browser tests against every live page, and a plain-English morning briefing waiting when I open the laptop.
• Quality is checked by machines, on a schedule. Accessibility and performance audits run automatically on every release. Every recurring failure gets logged in a reliability ledger and turned into a written rule, so the same mistake cannot happen silently twice.
• Nothing substantial ships without a structured review. Before major work goes out, it passes a standing multi-perspective review: strategy, legal, accessibility, marketing, and operations lenses. The blind-spot check a solo operator would otherwise skip.
• Research arrives as finished briefs. Competitive analysis, market sizing, and tool evaluations run as autonomous research jobs and land as sourced, decision-ready documents, not as open browser tabs.
• Incoming material sorts itself. A content-intake pipeline captures articles, ideas, and links as they arrive and triages them for review, so nothing depends on me remembering to file things.

The full working case study of how this is governed, what runs each night, and what the rules are lives here: how the system works →

The operator years

15+ years. Specific results.

How I work

The honest-operator rules

• Find the real problem first. Every engagement starts with framing, not tooling. Most "AI problems" turn out to be process problems wearing a costume.
• No hype, ever. If a tool is not worth adopting, I will say so, even when saying so ends the project. That is the point of hiring an independent evaluator.
• Bring the people along. Adoption fails on trust, not technology. 15+ years of leading teams taught me that nobody transforms an organization from a slide deck.
• Ship for the end user. The measure of the work is whether it genuinely helps the person at the end of the chain. Everything else is overhead.