Still an untapped opportunity in main stream application development for validation, early warnings, business workflow decisions, fraud detection and so on. Our applications may have these use cases mostly implemented by a data store, a predictive analytics or machine learning software and a visualization software. The usual data and information flow happens from a form based input from computer terminals with static validation. The input will trigger a business workflow with a rule based validation and finally ends in data store. Traditional application architectures try to establish above dataflow – a store and forward architecture. Once the above project goes into production (or in staging) phase, usually we see an opportunity of further exploring the data and start a business intelligence project. The transaction based data flow project and analytics based business intelligence project usually treated as separate cost model projects with a new team. The usual business intelligence project targets mainly on setting up a data warehouse, create OLAP models and render descriptive statistical or analytical reports. Creation and deployment of a statistical learning model is rarely happens from a traditional DW/BI project. Most of these projects are locked under the capabilities of SQL or SQL like technologies. The actual power of the data comes when it can use for prediction and prescription. The silo nature of predictive analytics project from BI and transaction projects are huge. The only interaction between the DW and PA team will be on data pumping. In this third stage of the data, from transaction application to descriptive analytics to predictive analytics, we can find an opportunity to make our whole transaction application suite to leverage the predictive analytics capability to make the application into an intelligent and self-learning one. Yes, we have seen the results of so called predictive programming in biggies apps and web sites. In very near future for most of the business applications, predictive programming will not be a nice-to-have but a must requirement. With advancements in tools and technologies around machine learning and big data, the opportunities for predictive programming is very promising.