Long-form essays on what it takes to put AI into production inside a regulated firm, and to stand behind what it produces.
The AI is ready. The trust architecture is not. Whoever builds it first wins.
Complementary intelligence: which parts of a process go to software, which to a model, and which stay with a person.
Why maker, checker, supervisor does not survive contact with machine output, and the shape that replaces it.
You need the judgment in their heads to build the system, and some of them will not be needed afterwards.
The framework pointed at one process: continuous assurance instead of a close, in fund administration.