E-commerce Giant Segments Customers in Messaging Apps Using Machine Learning SQL Database
Case Studies
Mar 11, 2023

E-commerce Giant Segments Customers in Messaging Apps Using Machine Learning SQL Database

The Challenge

An E-Commerce Giant's customer service team is working to improve the overall customer experience through their messaging channels. One of the most difficult challenges they face today is that the E-Commerce giant currently carries a wide range of products from numerous vendors, and each service agent has a limited knowledge of the product line. To improve the overall customer experience in messaging channels, the customer support team recognizes the need for a better way to segment customers from chat conversations and route the customer to the agent who can best assist.

The Solution

To address this issue, the customer service team is looking into ways to upgrade their messaging system to support customer segments. The messaging application was built on a NoSQL database prior to integration. Although the NoSQL database supports full text search, it lacks machine learning features such as text classification, which is essential for segmenting customers. However, developers can determine the sentiment of customers and the type of product category they are looking for by using the Superinsight ML database text classification feature. The most difficult part of this solution is determining how to build this new feature without interfering with existing features using an ML database.

To avoid interruptions, the development team decided to build this new feature on the Superinsight ML database while keeping all existing features on their existing NoSql database. The development team sent a duplicate copy of each incoming message to the ML table so that the messages could be queried by the new API they are creating. The development team was able to integrate the new Customer Segmentation feature with the existing messaging applications with no downtime after about 4 weeks of development.

The Result

Following the implementation of this new feature, customers were automatically labeled and segmented the moment they entered the messaging session. A month after the implementation, the customer service team was able to reduce support time by 20%. Following this improvement, the customer service team shares this customer data with the marketing team, allowing the marketing team to gain a better understanding of their customers for marketing purposes.

Nelson Chu
CEO and Co-Founder

He has over 20 years of experience in enterprise software and data management. Prior to Superinsight, he was in Sony where he worked on building a digital asset management system that takes analog video and moved them to the cloud. He was also in Disney where he worked on numerous data product projects.

Subscribe to our newsletter

Thanks for subscribing to our newsletter
Oops! Something went wrong while submitting the form.