AI is the next instrument of optimization

When I started, the instrument was the loop controller. Then it was the DCS, the historian, model predictive control. Each one didn't replace the engineer — it gave the engineer a new way to see the process and a new lever to pull. Performance went up each time, not because the work changed, but because the instruments got better.

I've come to think of AI the same way. It's the newest instrument of optimization. It doesn't change what operations is for — running better, more safely, at lower cost. It changes what's reachable. Things that used to be too unstructured to act on — maintenance logs, shift notes, the why behind a trip — become things a system can actually use.

AI isn't a pivot away from the work. It's the next instrument for doing the work I've always done.

The same caution applies

A new instrument is only as good as the discipline around it. In safety-critical work you never trust a new reading until you've verified it, and you always design for the failure mode. That instinct is exactly what non-deterministic AI systems need: humans in control of what matters, full audit trails, and safe failure modes so adopting AI never takes the operation down.

The companies that get value from AI won't be the ones with the flashiest demos. They'll be the ones who treat it as an instrument — calibrated, monitored, and held to the same standard as everything else on the floor.