Predictive Insights (Preview)
Predictive Insights (Preview) is a feature of Amazon Connect Customer Profiles that uses artificial intelligence to generate personalized product and content recommendations for your customers. By analyzing customer interaction data, Predictive Insights helps you provide more relevant experiences across all customer touchpoints.
How Predictive Insights works
Predictive Insights (preview) uses AI models to analyze customer behavior patterns and generate real-time recommendations. The service processes your customer interaction data, such as purchase history and browsing activity, to identify patterns and preferences.
-
Step 1: Add interaction data to profiles using existing data connectors to train AI models with your customer interaction data
-
Step 2: Add item catalog to S3 to allow Customer Profiles to access your item data through the AWS Management Console
-
Step 3: Create recommendations by defining recommendation types (similar items, frequently paired items, popular items)
-
Step 4: Apply recommendations across Amazon Connect ecosystem including Agent Workspace, Flows, and Connect AI agents
Prerequisites
-
Enable Data Store in Customer Profiles
To train AI models using your Customer Profiles, you need to enable data store. See details under Customer Profile Data Store to learn more.
-
KMS
You have configured Customer Profiles to encrypt your data under a AWS KMS key.
-
Security Profiles
You have configured Security Profiles to support View (list and view predictive insights), Create (create recommendations), Delete (delete recommendations), and Edit (update recommendations) permissions with Predictive insights enabled.
Benefits of using Predictive Insights
Using Predictive Insights provides several key benefits:
-
Improve customer experience with personalized recommendations
-
Increase sales opportunities through relevant product suggestions
-
Save agent time by automatically surfacing relevant recommendations
-
Deliver consistent recommendations across all customer touchpoints
-
Update suggestions in real time as customer behavior changes
Data Considerations
The following sections provide guidance on how to match use-cases and assess data readiness for Predictive Insights.
Have you matched your use cases to Predictive Insights?
Predictive Insights personalization types can address the following use cases:
-
Generating personalized recommendations for a user
-
Recommending similar or related items
-
Recommending trending or popular items
-
Re-ordering items by relevance
Do you have enough item interaction data?
For all use cases and personalization types, you must have at minimum 1,000 item interactions for 25 unique users with at least two interactions each. For quality recommendations, we recommend that you have at minimum 50,000 item interactions from at least 1,000 users with two or more item interactions each.
Do you have a real-time event streaming architecture in place?
If you have the ability to stream real-time events to Connect Customer Profiles, you will be able to take advantage of real-time personalization. With some personalization types, Predictive Insights can learn from your user’s most recent activity and update recommendations as they use your application.
Is your data optimized for Predictive Insights?
We recommend you check for the following in your data:
-
Check for missing values. We recommend that a minimum of 70% of your records have data for every attribute. We recommend columns that allow null values be at least 70% complete.
-
Fix any inaccuracies or issues in your data, such as inconsistent naming conventions, duplicate categories for an item, mismatched IDs across datasets, or duplicate IDs. These issues can negatively impact recommendations or lead to unexpected behavior. For example, you might have both “N/A” and “Not Applicable” in your data, but filter out recommendations based on only “N/A”. Items marked "Not Applicable" would not be removed by the filter.
-
If an item, user, or action can have multiple categories, such as a movie with multiple genres, combine the categorical values into one attribute and separate each value with the | operator. For example, a movie’s GENRES data might be Action | Adventure | Thriller.
-
Avoid having more than 1000 possible categories for a column (unless the column contains data for only filtering purposes).
