@bayareaguy
Smart scores and predictive scores were actually paid services packages. Our in-house data science team used to look at historical data to deliver sophisticated data-driven models. It is not a product feature. These packages are no longer being offered.
I agree that the documentation can be made clearer. We will make the appropriate changes. Thanks for bringing it to our attention!
@bayareaguy predictive scoring can be done in other tools for data analysis or you can build one using the rules engine. The key will be in defining what attributes lead to a healthy customer. Jackie from Valuize recently wrote an article that might be helpful in your quest to define your own predictive score.
https://valuize.co/blog/customer-health-score/