Before using AI segments, please be aware of the following: Prerequisites for AI features.
AI segments are driven by the Emarsys Artificial Intelligence learning algorithms that continuously train themselves by processing your customers' data and identify and analyze patterns of behavior. By using AI segments, you can predict the future behavior of your customers with high certainty based on their past behavior that allows you to encourage customer actions or prevent undesirable tendencies.
Have a look at a brief and portable summary of our AI capabilities along with some relevant customer testimonies.
How do AI segments work?
AI segments use prediction scores to calculate how likely it is that a contact will perform an action, for example, whether a customer will convert, remain inactive or churn. In general, prediction scores determine the thresholds based on which contacts are classified into 3 groups. AI segments provide you with ready-made templates, each one is represented as a single sentence containing various options that enable you to reach different target audiences. The following example shows you how the AI - Predictive lifecycle segments for leads template looks like:
Emarsys estimates the threshold values of the various groups behind the available options by using the statistical model of logistic regression. These thresholds are dynamically checked and updated by the Emarsys AI learning algorithms and vary from customer to customer.
Prediction scores, similarly to segments, are always refreshed before they are used, so when using AI segments, you will always work with the latest results. When you start using AI segments in your campaigns, your segments will become more precise in predicting customer behavior and the performance of your campaigns will increase significantly due to the Emarsys AI learning algorithms.
In addition, our hybrid infrastructure combines the Emarsys private cloud that meets the requirements of the highest security standards with public cloud services (Google BigQuery) that allow us to make performance-heavy calculations. This robust infrastructure enables us to carry out complex estimations securely and rapidly.
Why should I be using AI segments?
With AI segments you can:
- Target your contacts with emails having different subject lines or content based on your customers' engagement levels.
- Spare money on Digital Ads by targeting only those contacts who are likely to remain inactive or disengage or by targeting these contacts on different channels.
- Lead your customers from one lifecycle stage to another more precisely and efficiently, for example, you can encourage your first-time buyers to become active buyers.
- Differentiate your leads who are likely to convert from those who are likely to remain inactive or become cold.
- Distinguish your customers who are likely to engage with your campaigns from those who are likely to remain active or disengage.
Available AI segment templates and options
The following table lists the available AI segment templates and options.
Template name | Description | Available options |
---|---|---|
AI – Predictive email engagement | You can target contacts based on their expected reactions to your email campaigns with this template. | Contacts who, in the next 30 days, are likely to:
|
AI – Predictive web engagement | You can target contacts based on their expected reactions to your web campaigns with this template. | Contacts who, in the next 30 days, are likely to:
|
AI – Predictive lifecycle segments for customers | You can target contacts based on their buyer status and differentiate churning customers from those who are likely to make a repeat purchase with this template. | You can target the following contacts:
and who, in the next 30 days, are likely to:
|
AI – Predictive lifecycle segments for leads | You can differentiate leads who are likely to convert from those who will possibly become cold and target them with this template. | Leads who, in the next 30 days, are likely to:
|
AI – Predictive revenue | You can target contacts whose next cart value is likely to fall in the bottom / mid / top third of all your contacts. | Contacts whose cart value is low / medium / high. Note: AI segments determine the cart value of a contact by analyzing the cart values of all your customers. |
All of the AI Segment templates leverage both Smart Insight and Web Extend data history. The only exception is AI Predictive revenue which solely relies on Smart Insight data exclusively.
Note that the First-time buyer label at AI Segments is different from the Smart Insight First-time buyer. The same contact can be categorized differently, as Smart Insight does not take Web Extend into consideration at calculation.
Creating an AI segment
To create an AI segment, proceed as follows:
- Go to Contacts > Segments, click Create Segment, select Standard Segment, then click AI segment.
- Give your segment a Name and add a Description to it.
This can help you identify segments easily on the Overview page.
- Select a template and choose the option you need.
- To create and save the segment and display the results, click Save & Apply.
Notes:
- Segment results are cached for 10 minutes. So, if you run the same segment again within 10 minutes, you will get the previous (cached) results.
- You can use AI segments independently of the segments in Automation Center programs or on other channels, too.
Optionally, you can combine your AI segments with different segment types by clicking Combine this segment. For more information, see Creating combined segments.