Operators, Not Evangelists
Somewhere in the past year the question changed. The rooms I sit in used to ask what AI could do. Now they ask something harder: who can make it work here — inside this infrastructure, this compliance envelope, this team, this book of business. Those are different questions, and they call for different people.
An industry newsletter that landed in my inbox this week told the story of a CEO who offered to pay, out of her own budget, to fully train a new hire on AI. The catch: she was not looking for an AI expert. She kept meeting candidates who were fluent in the tools and lost in the business, and she had concluded that tool fluency is trainable while operational judgment is not. I had been circling the same thought for months. It was strange, and a little confirming, to see it arrive from someone else’s desk.
The market is splitting
On one side are the evangelists — people who can present AI beautifully, demo impressively, and leave the room without anything changing. On the other side are seasoned operators who know the business cold but treat AI as a topic for a slide rather than a material they can work with. The value sits in the overlap: someone who can hold the operational problem and the engineering reality in the same head. That overlap is nearly empty, and organizations are starting to notice.
What the overlap looks like
I can describe it concretely, because it is where I work. It looks like knowing carrier messaging compliance well enough to get a text-message lookup tool registered and approved — and knowing brokers well enough to know why that tool should exist at all. It looks like walking an AI-written application through a HIPAA security review — the briefs, the risk assessment, the sign-off gates — because a real customer relationship is waiting on the other side. It looks like twenty years of operations instinct deciding where AI belongs in an existing process, and just as often, where it does not.
None of that is glamorous. All of it is the job. The tools were never the hard part; the hard part is the seam between the tool and the organization — the compliance envelope, the data agreements, the handoffs between people and systems, the process that has to keep running while you change it.
Pulling the pieces together
The call I keep hearing — from executives, from operators, from the market itself — is not for more AI presentations. It is for people who can pull the pieces together: who understand how AI gets engineered into an existing infrastructure, framework, and process, often incrementally, to produce results a CFO can see. Not AI as the answer. AI as a component, engineered in, with the operational thinking that makes it hold.
If the last two years were about convincing organizations that AI matters, the next ten belong to the people who can make it matter operationally. That is the position I have chosen on purpose: one foot in the engine room, one foot in the market, and a working belief that the distance between a strategy slide and a running system is the most valuable ground in the industry.