Estimated reading time: 4 minutes

The Chemical Connection: Employee and Operator Empowerment in the Age of AI

In an ideal world, every operator would understand the process they are working on and be able to diagnose problems and suggest improvements. In the real world, however, we all know that these types of people are, if not rare, at least in the minority. Where do these people fit in when everyone is turning to automation and AI for quality and yield improvements and cost reduction? (Never lose sight of the fact that, in the end, it’s always about the money.) In my opinion, these people are valuable resources, and every effort should be made to retain them, regardless of their education level or seniority.

During my career, I’ve had the privilege (mostly) of visiting dozens of PCB manufacturing sites, good, bad, or indifferent, and have formed the opinion that how these facilities treat their empowered operators is a sign of how good or bad that facility is. Here are some examples.

Examples of the ‘Good’
Just a few months ago, I participated in a video conference with a smaller shop that was looking to add some automation to its cupric etch system. They had a basic etcher that depended on an experienced operator to keep the system running. He was the type of guy who was continuously monitoring the specific gravity, ORP, and free acid levels, etc., to make sure they stayed within the optimal operating parameters and was trusted enough to do this with minimal supervision.

That operator was in the video conference along with the process engineer and plant manager, which, to me, was a good sign of a good shop. After we had come up with a system that met their needs and price range, I pointed to the operator and said, “Whatever you do, don’t get rid of this guy. The automated system is good, but it still needs direct supervision.” The manager said, “Oh no, the plan is that, since he doesn’t need to be constantly monitoring the etch chemistry, we promote him to supervising the whole DES operation so our process engineer can spend less time there and work on other problems.” That, in my opinion, is exactly how it should work.

Another time, I went to a facility to help with a process problem and was left to solve the problem on my own while the process engineer went off to do something else. This happens far too often, but in this case, the etcher operator took an interest in what was going on and asked all kinds of questions about what I was doing and why. By the time the problem was resolved, he knew exactly what had happened and took responsibility to see that it never happened again.

I told his supervisor to give this guy whatever he wanted in terms of wages because lost time and production while trying to replace him (if they could find someone as good) would be more than any salary increase. The supervisor smiled and said he knew that and that it wouldn’t be long before he was running the entire etching department. When I visited again six months later, that’s exactly what had happened. Again, exactly how it should work.

An Example of the ‘Bad’
However, the idea of using automation to free up time for a competent operator or process engineer to expand their responsibilities can be taken too far. Several years ago, I visited a rather large PCB shop on the West Coast to help with a process problem involving our equipment. I met their process engineer, who was responsible for the entire wet processing (DES, SES, plating, etc.) area across the whole facility.

The actual problem turned out to be a rather simple one with an easy fix, but he was constantly being called off to other sections of the plant to deal with other issues, so it took several extra hours to find and fix it. For months after that visit, I got phone calls and emails from him, time-stamped 10 p.m. or 3 a.m. and other ungodly hours. He was always at the plant for one emergency or another.

About a year and a half later, I got a call from a headhunter offering a wonderful job on the West Coast at a generous salary increase. As she described the job, I began to get suspicious and asked if the job was at this company I had just visited the year before. Well, yes, it was.

I politely turned down the opportunity since I wasn’t eager for a job that required me to be on call 24 hours a day, seven days a week (that and the fact that the rent on an apartment a one-hour commute from the plant was more than the mortgage on my house, a 15-minute commute to the plant, in State College, Pennsylvania). It isn’t always about the money, you know. I later learned that the process engineer had burned out, suffered a nervous breakdown, and left the company to recover and seek sunnier climes.

So, using more automation to free up time for good employees to expand their responsibilities and retain their services is a good idea, but don’t take it too far.

And What About AI? A History Lesson
Now we come to AI. I freely admit that I am uneasy about it—not so much about AI itself as about the motives of the people rushing to embrace it. It seems to me that the underlying (and unspoken) reason AI is so eagerly anticipated by some isn’t that it can increase efficiency and quality (which it can), but for the potential to replace employees that need to be paid and thus maintain or increase profit levels. Granted, this will happen, but where do the reductions in personnel come from?

Back in the ancient times of the PCB industry (1980s), American PCB shops steadily lost business to Japan, which supplied high-quality circuit boards at very competitive prices. The Japanese had aggressively automated their production and run their plants with a few well-trained and well-paid employees. The response from many American shops was to lay off their highest-paid employees and hire more minimum-wage workers to cut costs and compete on price. Most of the shops that adopted that tactic were soon out of business, and the Japanese shops took over even more of the business.

AI is coming, of that there is no doubt. My observation is that the good shops will use AI to increase the efficiency and productivity of their empowered employees, not to get rid of them. In my opinion, based on reports of experimental testing and real-world failures, no AI should be allowed to run without some human oversight and supervision.

This column originally appeared in the September 2026 issue of I-Connect007 Magazine.