When Should Teams Consider Building Their Continuous ML?

The first time I came across the maturity levels of ML engineering platforms is done a Google paper from four or five years ago. Now you have the level zero, they call it, where everything is super manual. The last level and to be honest, I'm meeting many companies these days and I'm trying to figure out how they're utilizing their ML infrastructure. So maybe this is like the level three, the level two or level three for those maturity levels in organizations.