Insights
Writing
Notes on industrial AI, decision intelligence, and building practical AI for the people running heavy industry.
AI Won’t Run Your Cement Plant. Here’s What It Actually Does
Every other pitch promises AI will run your plant. It will not, and you should not want it to. Here is the honest map of where it actually helps across the operation — and where it cannot.
Cement’s Most Certain Risk Isn’t Carbon — It’s Demographics
The industry spends its strategy on carbon and capital. The most certain risk over the next decade is neither — it is demographic, it is already underway, and you cannot hire your way out of it.
When the Expert Retires: Cement’s Invisible Safety Risk
The kiln, the dust, the confined spaces — we assess those. The risk that is not on any hazard list is the expertise retiring faster than we capture it. On cement’s real safety problem.
Why Industrial AI Pilots Die on the Plant Floor
A slick AI dashboard, six months later, has four logins — all from the same office, none from a control room. The pilot gets quietly declared inconclusive. The autopsy almost never finds a technical cause.
What I Learned Writing a White Paper on MSHA Compliance
I set out to write about MSHA compliance and kept landing on the same thing: most of it is institutional knowledge, not legal complexity — and that is exactly what walks out the door at retirement.
Why Cement Needs Domain-Specific AI, Not Generic AI
Free lime at 6 a.m., a guarding finding, a mill down at 3 a.m. — three questions that show exactly where generic AI fails on a plant floor, and what domain-aware decision support does instead.
From the Plant Floor to Industrial AI
A reading drifts past midnight, and the one person who knew the cause retired last spring. That night — repeated for years — is why I left the lab bench to build industrial AI.