The problem is, the shit that's made up will be used to justify the decision as a success and ensure the methodology continues.

If they’re mandating use like this I doubt it’s their only dysfunction. At least this one has a built in scapegoat.

Sounds like they are going to consider it a success no matter what.

Make it long enough that it's not worth any human's time reading it. Like full-on balls to the walls slop. Pass it between different LLMs to add additional conclusions and subheadings.

In fact, do it in parallel where one chatbot is adding another few pages here, and simultaneously and independently, another is adding different pages somewhere else and concatenate the results together.

Once you get about 25 pages of dense slop, just conclude that AI made writing this report 1000x more efficient.

Their ability to convince themselves they are geniuses far exceeds your ability to convince them they're not. They've already decided that the decision was a success. The only question is are you still going to be around to suffer the consequences when their delusions collide with objective reality?