Nice and well written. Gives so much information. Appreciate for publishing it. The point is valid that instead of delivering model at the end of 1 year, regular updates should be given by ML team every month so that there is no surprise when the model is developed and delivered. I was given a task to identify the reason for a project which was delayed by 1 year and for which delivery was done but not what the customer had given requirement for. One more point I want to add in your article that proper documentation like BRDs/ FSs are also very important on which all stakeholders should agree.