Nowadays businesses have too much customer data for a human to handle. Sifting through heaps of it manually trying to find insights is not only gruelling, sometimes it's no longer possible. Solution: let the self-learning algorithms find patterns in data that yield business value. We'll show how a major online retailer, a DIY store and an airline booking service cooked a conversion prediction system out of their various sources of data – CRM, web, and app. This won't hurt, almost anyone can do it – and we'll demostrate how.