OP here. This is something I've been pondering recently and have seen to some extent in all companies that I have worked with. I've found that objective discussions, based on mutually agreed pros & cons, are often more important than the specific machine-learning model used.