Traditional predictive analytics have a solid grounding in statistical modelling and allow a business to analyze its current and historical data in order to identify opportunities and risks for the business going forward.
Machine Learning can be thought of a branch of artificial inelligence that allows computers to infer rules from business data without having to be explicitly programmed. It has typically required large quantities of data, expensive computing resources and tools which require very specialist knowledge to produce sensible answers.
As these two disciplines converge and the cost of computer power declines, it is time to consider how can we make best use of this data without an army of data scientists and a huge budget. This is becoming increasingly important in a world where we have considerable amounts of digital business data which often isn't stored in logcal and easily accessible ways.