Sample use cases

Explore some of the ways you can experiment with AI/ML analytics on large datasets, before scaling, applying the results to other use cases, or operationalizing your models.

Tip

For help with these use cases, email the support team or ask the community.

Before you begin

• Use the Project Administration notebook to prepare your project.

• Use your existing AWS Glue catalog and upload our sample data. Or create a new catalog to use the provided Open Table Format (OTF) examples.

Understand the customer journey

Customers engage with financial institutions through various marketing touchpoints including websites, in-branch interactions with employees, email, and call centers.

Understanding the overall customer journey is critical for enhancing it and increasing customer acquisition and adoption.

After you access and inspect the data, you'll use attribution modeling, channel analysis, and pathing to understand the customer behavior over time.

Try it out (SQL)

Try it out (SQL-Python)

Segment customers based on past purchases

Purchase history can be a powerful factor for segmenting customers. For instance, you might create segments based on purchase volume and value.

After you prepare the data, you'll use k-means clustering and data preparation pipelines to determine customer segments.

Try it out (SQL)

Try it out (SQL-Python)

Find the ideal number of customer segments

Marketing to the ideal number of customer segments means you benefit from segmentation without having too many segments to manage. You can derive the ideal number of segments from comments from customer purchase history.

After you access and inspect the data, you'll use a large language model (LLM) to create a vector table, and then find the ideal k-means model and number of segments.

Try it out (SQL)

Try it out (SQL-Python)