This is the area I struggle most with, in terms of predicting the future. Although it's not like any of us can be great at forecasting the future in the first place.
But: if I were given complete control to build a department from absolute scratch, I could definitely deploy tools and platforms and regulations/policies in such a way that I could be the solo doc staffing a very busy practice that spans multiple sites, operating 5 days a week.
The issue is...that world doesn't exist. Right now, for example, I'm several months into trying to "fix" a department that was ignored for years and years. I have almost cheerleader-level support from admin - not quite "blank check" but...close.
Even with this setup, where many key admin/executives are backing me to a weird degree, and I have the knowledge and ability to drag this department into the modern era - the systems and inertia AROUND this department mean this will likely take me several years.
I'm very "techno optimist" for AI in RadOnc, but I recognize I might be wrong. My optimism comes from me being in places where I'm the only one that really understands it, and can use it to make myself more efficient.
One day I'll encounter an executive who understands what is possible on a high level.
Perhaps I'll feel differently after that happens.