AI model deployment is more than just model scoring. A good-performing AI model does not guarantee a successful deployment. You will need additional analytics.
Though there are many technical check-list exists for a AI deployment, I will focus on a checklist from business perspective.
Business Process contextualization of AI model scoring
Let us assume that you have a churn prediction model for telecommunication customers. You need to think about the business process in which the model will be deployed. A churn prediction model could be deployed in the various business processes. Take two scenarios :
Automated customer billing process: The objective is to provide a discount to customers who have a high probability to churn
Customer care call-center process: The objective is to provide a visual understanding to the call-center operators on churn analysis for the customer. If…
