SEGA Europe builds a player sentiment analysis GenAI model and Genie One self-serve analytics on Databricks

Overview

SEGA Europe consolidated data from over 40 million players generating 50,000+ events per second into Delta Lake with Lakehouse Federation and Unity Catalog, then built a generative AI player sentiment analysis model that analyzes 10,000+ user reviews daily to identify gameplay issues, and deployed Genie One for natural-language self-serve analytics, achieving a 10x faster time to insight.

The challenge

SEGA Europe previously grappled with extensive amounts of data and data sources, with over 50,000 events per second from more than 40 million players across more than 100 video games, struggling with data integration, quality, and accessibility, and eliminating data silos across marketing, sales and game development teams.

The solution

SEGA Europe integrated all its data into Delta Lake on the Databricks Data + AI Platform, using Lakehouse Federation to query disparate databases including Redshift, BigQuery and SQL Server, governed by Unity Catalog for a single source of truth. Databricks SQL and Genie One let users across departments access data in plain English, AutoML let teams try out machine learning use cases within minutes, and the team built a player sentiment analysis generative AI model that translates and analyzes tens of thousands of user reviews daily.

Reported business value

SEGA Europe achieved a 10x faster time to insight for business users with Genie One driving self-serve productivity, and its player sentiment analysis GenAI model analyzes 10,000+ user reviews daily to pinpoint and address game-related issues, significantly increasing player retention in certain SEGA titles.