Exercise 4: Create a Predictive Model

In this Exercise, you will create and train a predictive model to help predict the lead conversion rates of a new business.

Step 1: Create an AI Model to Predict Lead Conversion Probability

Create a predictive AI model to predict lead conversion probability based on historical data.

1. In the Data Cloud app, select the Einstein Studio tab.

2. Click Add a Predictive Model.

3. Select Create a Model from Scratch.

4. Click Next.

5. Under What data would you like to use?, select Lead Engagement Training Data.

6. Click Next.

7. Click Next.

8. Set the following training values:

   Field	Value
   Which field would you like the model to predict values for?	IsConverted
   Would you like to maximize or minimize IsConverted?	Maximize
   isConverted	TRUE

9. Click Next.

10. Set Autopilot to Disabled.

11. Select the following fields for training:

    • TotalIndividualRecords
    • TotalWebEngagements
    • Source
    • WatchedVideo
    • EngagedWithEnterpriseElements
    • SubmittedLeadForm

    

12. Click Next.

13. Click Next.

14. Click Save.

15. Change the model name to Predict Lead Conversion Probability.

16. Click Save & Train.

TIP

Predictive models can take up to 24 hours to train. With the data set that we have selected, it has taken an average of 20 minutes to train, but may take longer due to the size of the workshop.

Summary

You now have a model that is custom to your org being trained in the background. We will be able to use this model across the platform to gain insights and make predictions.