In my late thirties I was C-suite at a tech company, earning well, and driving around Cape Town in a Mercedes I could very much afford. I would sit at the traffic lights and look at the people in the cars around me and think, quite clearly: I earn more than them, so obviously I am better than them. I'm not ashamed of this. I'm more impressed that I can look back and catch myself in the middle of it, a woman in her late thirties, in an expensive car, in a beautiful city, doing

Professionalism was never a description of competence. It was a set of visible proxies for trustworthiness, assembled alongside the expansion of Western bureaucracies, corporations, universities, law and medicine in the nineteenth and twentieth centuries. The suit. The controlled voice. The correct accent. The restrained body. The office. The separation between public self and private self. Two models fused into what we now call professional. The American corporate man, sober

Most AI deployments in insurance get the technical implementation right and the human integration wrong. The model performs. The underwriters and claims handlers route around it, override it, or quietly stop trusting it. Six months in, the productivity case doesn't hold and nobody can pinpoint why. The industry is measuring the wrong thing. Model accuracy, latency, cost per inference - all tracked, all dashboarded, all beside the point. The actual failure is happening in the