Recently, a whole host of product updates were put into place by Oracle, specifically to the EPM platform. Most importantly, the long awaited Auto Predict functionality, which was initially demoed in the August 2020 release, is finally available for EPM standard and EPM Enterprise applications. Predictive Planning remains one of the core competencies for the Promethean Team, and we’ve been eagerly awaiting this addition to our toolkit. 

Let’s dive in.

A common issue that clients dealt with was the limitations of the last generation Predictor Engine. Namely, the cumbersome process that involved pulling together past sales data, adding exceptions, and executing the process via manual triggers. For clients that were managing hundreds if not thousands of SKUs/departments/etc., the forecast became limited to the historicals and bogged teams down with numerous manual adjustments. To some extent, the prediction results were only as accurate as the historical data you had, with any gaps in information existing as glaring holes that the engine couldn’t paper over. This led not just to time lost, but hours if not weeks of valuable time spent repairing the issues in data and the exceptions, as opposed to evaluating results and charting a pathway forward. Any unexpected changes to the model or the data, then, becomes a huge time and energy sink.

Auto Predict does more than fix the missing links in the data, it opens the door to a whole new level of planning and forecasting for clients. Available for all hybrid enabled EPBCS Planning Apps, it improves the prediction process and user experience in multiple ways. Forecasts can now be approached from a goal oriented perspective than as a task, and crucially opens the doors for clients to embrace variance exception based forecasting. Users can now identify the exceptions to their data that trigger the different results and build fresh forecasts with changes in those exceptions. Forecasts and planning scenarios can be seeded with exceptions of all types and allow teams to parse the biases in their model alongside the validity and strength of their historical data. The more data put in the better, and Auto Predict is great for large amounts of data (read: SKUs). From there, the data is crunched through 13 advanced statistical models and uses the best fit method for the right prediction. 3 Prediction versions can be generated for the client, namely Base Prediction, Best Case, and Worst Case. With the ability to understand the variances making the differences, teams can gain a whole new vision into their data and the future. Bottom line: More accurate forecasting, here we come.

Even better, the process is now automated and extensible through Rest API, EPM Automate, and Groovy, and can be scheduled as a job to be run as any client sees fit. This allows plans and forecasts to be generated with the freshest possible data as often as needed, allowing customers to receive the most accurate data when they need it. Planning and forecasting cycles can be dramatically sped up, allowing resources to pivot away from data collection / repair and instead be focused on in-depth analysis and decision making. It’s not just better forecasting, it’s creating a better model for teams to trust and rely on for key decision making.

Promethean is extremely excited about this new enhancement to to predictive modeling.  By the addition of this feature, it places EPBCS miles ahead of the competition.  Not only does predictive modeling provide advanced modeling capabilities on predicting future patterns, but now we have the ability to automate this process at will.  It takes our included predictive capabilities, of our Accelerators, to the next level. Gerry Villamil, Managing Director, Promethean Analytics

This is just a bit of what Auto Predict can do for you and your planning team, and we’re happy to help you get there, either through our powerful Solutions or Accelerator programs.. For more information or a demo, reach out to