Case Study

Word Collections

Helping Music Rights Teams Automate Licensed Content Detection at Scale

The Client

Word Collections is a global music publishing administration and royalty management company founded by Jeff Price, the industry innovator behind TuneCore and Audiam. The company helps songwriters, publishers, and rights holders maximize royalty earnings through modern rights management, royalty tracking, and licensing solutions.

As the volume of newly released music continued to scale globally, Word Collections needed a more automated and intelligent way to identify derivative works, licensed recordings, covers, and reused content across massive streaming datasets.

RevStar partnered with Word Collections to productionize an AI-powered audio waveform identification platform capable of continuously analyzing newly released music and detecting royalty-eligible content at scale.

Word Collectinons transparent logo
Cloud-Cost-Optimization

Summary

TYPE OF PRODUCT

AI-Powered Audio Identification Engine

BUSINESS VERTICAL

Music Publishing & Royalty Management

TECH USED

AWS SageMaker + Machine Learning + Cloud-Native Architecture + Data Lake Infrastructure

SERVICE
Full-service Development Team—

Solution Architecture, Delivery Management, AI/Data Engineering, DevOps Engineering, Quality Assurance

DESCRIPTION

The solution exceeded 90% accuracy benchmarks, reduced projected manual review effort by more than 50%, and established a scalable foundation for continuous AI-driven royalty recovery operations.

RESULTS

Enabled faster and more scalable royalty recovery operations through automated licensed content detection and AI-powered audio analysis.

We are very impressed with RevStar. Their team was able to very rapidly come up to speed with our complex business requirements, and synthesize and implement effective solutions within AWS.
David Willen President & CTO, Word Collections

The Challenge

The rapid growth of global music streaming created a significant operational challenge for Word Collections’ royalty management workflows.

Manually identifying derivative works, licensed recordings, covers, and reused audio content across monthly streaming reports became increasingly time-consuming, inconsistent, and difficult to scale. As new releases accelerated globally, human review processes could no longer keep pace with the growing volume of audio requiring evaluation.

Without an automated system for continuous analysis, delayed identification of royalty-eligible recordings increased the risk of missed claims, delayed revenue recovery, and operational inefficiencies.

Word Collections needed a scalable, production-grade solution capable of continuously analyzing newly released music with high accuracy while minimizing manual intervention.

Platform-Thinking-2026

The Problem

As global music releases continued to grow, Word Collections needed a more scalable and automated way to continuously identify licensed audio content across large streaming datasets while reducing reliance on manual review workflows.

The Solution

We designed and deployed a secure, AWS-native licensed content detection platform powered by machine learning and scalable cloud infrastructure.

The solution automated the analysis of newly released audio content, helping streamline royalty identification workflows while reducing the operational burden of manual review. Built to support rapidly growing streaming volumes, the platform was designed with scalability, reliability, and long-term operational efficiency in mind.

To support ongoing growth and continuous improvement, we implemented cloud-native infrastructure, automated processing workflows, monitoring systems, and operational tooling aligned to AWS best practices for security and performance.

What We Built

To meet project requirements we built a production-ready AWS-native licensed content detection platform powered by TensorFlow and AWS SageMaker to automate large-scale audio analysis and royalty identification workflows.

To support growing global streaming volumes, RevStar designed and deployed scalable machine learning infrastructure capable of continuously analyzing newly released audio against large licensed content datasets. The solution combined automated processing workflows, secure cloud-native architecture, and operational monitoring systems to improve scalability, reduce manual review requirements, and support long-term AI-driven rights management operations.

The Team

We have a growing cost-effective hybrid team, with onshore Product Managers and developers from our build center in Colombia. Operating three parallel teams for different products and support.
Two people collaborating at a whiteboard covered with sketches and sticky notes, planning a project layout.