Skip to main content

Celonis Product Documentation

The right tools for the job
Deploying your model

After you developed your machine learning model it is time for them to go to production. After all, spending all the time on building a good model has to lead to results!

Let's say you have a model that can predict if an order will be delivered late. You might want to have fresh predictions every day in order to act quickly.


Using the scheduling functionality of ML Workbench, you can trigger your ML model to calculate new predictions on a scheduled basis. In our example, you could calculate the predictions every night so the new insights can be used first thing in the morning.

Operationalize your predictions using Action Engine

A prediction itself is not worth anything. It is also important to actually act them! Using the Action Engine, you can configure Signals that, in our example, would inform the responsible employees that a specific order is late. The user can then directly prioritize the shipment from the Action Engine Signal.