My personal opinions, essays, and editorials on AI and where it's headed — written in plain language, straight from the work. Strong views, loosely held.
Most AI content is written by people who watch. This column is written from inside the work: two decades in enterprise tech, SRE and cloud architecture at Google, machine learning at BMW of North America, big-data marketing science at Fashion Nova, and now a Fractional Chief of AI seat inside owner-led businesses doing roughly $5M to $50M a year.
Every piece here traces back to something I actually built or broke. The voice AI essay exists because my agents answered 50,000 calls for a real client. The ROI essay exists because I kept watching businesses pay for dashboards that moved nothing. No roundups, no hot takes on somebody else's demo. Receipts or it doesn't run.
New here? Start with The Fractional Chief of AI, then Measuring AI ROI, then Voice AI That Actually Converts. Those three cover what I do, how I measure it, and what it looks like shipped. Evaluating the role itself? The 2026 buyer's guide has the costs, the comparisons, and the questions to ask. New essays land when they are worth your time, not on a content calendar. The 90-Day Memo at the bottom of this page puts one field-tested play in your inbox between essays.
Field data from a mass-tort reply engine — what makes an AI phone call get answered, trusted, and filed.
Read ↗The exact free and near-free tools I used to lift a Wynwood restaurant's profit 25% with AI imagery.
Read ↗What happens when you hand frontier tools to kids the industry forgot — and why it's the most important work I do.
Read ↗Reliability is a product feature. The same discipline that keeps planet-scale systems alive is what keeps AI in production.
Read ↗Most AI dashboards measure activity, not outcomes. Here's the small set of numbers leadership should actually track.
Read ↗Latency, fallbacks, and guardrails — the structure I use so a voice agent holds up under real call volume.
Read ↗The most important AI work isn't at the frontier labs. It's in the neighborhoods deciding who gets to build.
Read ↗Not the next Silicon Valley — something more interesting. Builders shipping AI for real businesses, today.
Read ↗Big-data marketing science taught me where AI cuts cost without cutting reach. The playbook transfers further than you'd think.
Read ↗The 90-Day Memo: what I'm actually building and what it moved, in plain language. No spam, no hype — just the work, when it's worth sharing.