Introduction
Model automation enables you to continuously use the model's predictions on your data and integrate these predictions with your business applications to embed them in your daily business operations.
You have two options on how to receive the prediction outcome:
Using a CSV report via email
Pushing the predictions back to one of the connected data sources. The platform will update the following values for each prediction:
Score - the prediction's score
Explain - the explanation to why this score was given
Trend - scoring trend (up, down or no change)
Scoring date - the date of prediction
Set A New Automation
Go to the model you would like to automate and click on the 'Automation' tab
Click on 'New' to create a new automation
For each automation, you'll need to define 3 parameters:
Read from:
The 'Read from' configuration enables you to define additional conditions on the audience you would like to filter, for example, only predict leads that were created in the past day.
Update to:
The 'Update to' defines how to deliver the prediction output. It could be sent as a CSV report by email or integrated back into the data source, as in the example below.
Update source - either by email or to each connected data source.
Object - If you want to push the results back to the data source, define which object should be updated.
Record ID - which key should the platform use to match records.
Score | Explain | Trend | Scoring data - the field that should be updated for each set. We recommend creating dedicated fields.
Sync:
This is the automation scheduler, for example, every day at 12 am. You can add multiple syncs by clicking on 'Add'.
Turn the automation on by clicking on the off/on toggle.
