Ask HN: Office politics decide ml model deployment

This is funny but there are so much hype around deep learning that managers use it as a tool to do politics. I have seen in multiple occasions the model is deployed on the basis of office politics rather than need or accuracy. What are your experience?