Transforming media monitoring with intelligent agents and metadata automation
Revolutionising media intelligence through hybrid search, automated metadata normalisation, and proactive quality control.
SoundAware
SoundAware and RTV Monitor required a scalable solution to handle massive volumes of broadcast data, including TV transcripts and music metadata. The project aimed to replace manual operator tasks with intelligent AI agents capable of hybrid search, complex metadata merging, and automated quality reporting for 146 channels. The goal was to improve operational efficiency while ensuring absolute accuracy for international reporting obligations.
Managing 800 hours of daily broadcast content and millions of music tracks created significant bottlenecks for operators. Manual metadata merging from diverse suppliers was prone to error and difficult to scale across 300 million records. Furthermore, verifying monthly reports for German broadcasters was a high-risk, manual process that often resulted in a three-month lag between broadcasting and reporting, creating significant financial delays.
A look inside
What SevenLab built
SevenLab developed a unified platform featuring three core AI-driven modules: a hybrid search engine for transcripts, a three-layer metadata normalisation pipeline, and an automated XML reporting validator. This system utilises PostgreSQL with pgvector for semantic retrieval, Mastra for agent orchestration, and LLMs for complex data reconciliation and automated summarisation.
Hybrid retrieval engine
Combining lexical and semantic search to allow precise transcript analysis and natural language questioning.
Three-layer normalisation
A deterministic, embedding-based, and LLM-driven pipeline for merging music metadata from multiple suppliers.
Automated quality control
Proactive detection of EPG gaps, coverage issues, and anomalies in monthly XML reports using AI-driven analysis.
Version-controlled metadata
A Git-style history for all metadata changes, ensuring full auditability and the ability to revert to previous states.
Measurable business impact
The implementation of the AI-driven suite has transformed operational workflows, enabling the processing of up to 40,000 tracks daily with minimal human intervention. By automating the pre-check of XML reports and the normalisation of supplier metadata, the agency has significantly reduced its operational risk and accelerated the reporting cycle. The hybrid search tool has also improved client retention by allowing users to extract deep insights from TV transcripts instantly.
The platform has fundamentally changed how we handle broadcast data. By automating the pre-check of our monthly reports and the merging of complex metadata, we have significantly reduced our operational risk and improved the quality of our service for major broadcasters.
Want results
like these?
Tell us your challenge and we'll show you how we'd solve it — with a clear scope, timeline, and fixed price.
Talk directly with our AI specialists