Schneider Electric, which makes electrical equipment for data centers and factories, has built an AI curriculum with four levels. Every employee takes the basic one. Senior managers get a course on AI and strategy, the engineers who write AI software get a technical one, and the people who lead the company's change projects get their own.
The mandatory course for everyone is the least interesting part of that, and the history of electricity shows why.
A new tool placed inside the old layout changes little
In the first wave of electrification, factory owners swapped the steam engine for an electric motor and kept everything else. A single power source still turned a shaft that ran the length of the building, and belts carried the power down to each machine. Because power was lost over distance, the machines had to stay close together, and the floor was arranged around the shaft, not around the work. The owners got a modest saving on fuel and nothing more.
The big change came later, when new factories gave every machine its own motor. Without the shaft, a factory could be a bright single floor, arranged to follow the flow of materials, and it could be rearranged whenever the product changed. The motor was the same in both cases. The difference was that someone rebuilt the process around it.
AI at work is at the motor-swap stage in most companies. A basic course teaches people to paste an email into a chatbot and get a summary, and that saves a few minutes in a process that is otherwise unchanged.
The person who knows the work has to be in the room
Take a company that sends repair technicians to customer sites, a job where Schneider uses an AI tool for dispatching to assign the work. Every morning a dispatcher matches the day's requests to the people available. She knows that one customer needs a certified technician, that another site requires two people for safety, and that a certain part is always out of stock on Mondays. None of that is written down.
After a course, she can ask a chatbot to summarize the morning's requests. After a project, she and an AI engineer and someone from IT have turned her rules into a tool that proposes the day's routes, and she only decides the exceptions. The engineer could not have written those rules, and she could not have built the tool. That is the logic behind mixed teams of business, IT and AI people, who learn the tools by building something together.
What to watch for
The fourth group in Schneider's curriculum, the people who own processes and run transformation, is the one that decides which layouts change. The mandatory course is the visible part, and a company can count completions and feel finished. Completion counts measure motor swaps. The number that shows real change is how many processes were rebuilt, and I expect the companies that track that one to pull ahead of the ones that track the first.
The senior-management course matters for a related reason. Most staff don't know why their company uses AI, and managers who can explain the reason are the ones who can ask the right people to redo the work.
If you run a small business, you can skip the course for everyone. Pick one task that repeats every week, such as reordering stock or answering the same customer questions. Put the person who does it together with one person who is comfortable with AI tools, and rebuild the task from start to finish. Then pick the next one.