Telecommunications, Media & Entertainment · Asia · sports broadcaster · Generative AI Solutions
A major sports broadcaster in Asia (kept anonymous) wanted to transform how live sports moments are captured, edited, and shared, aiming for smarter, faster, more personalized viewer experiences. Three obstacles stood in its editorial workflow: reliance on third-party providers (rigid, less customizable, misaligned with editorial perspective); slow turnaround (manual editing and delayed data feeds limited real-time responsiveness); and evolving fan expectations (personalized recaps, mobile-ready highlights, interactive catch-up features). The broadcaster needed to internalize control of its data and deliver highlights instantly without compromising creativity or editorial independence.
The team launched an internal "Open Innovation" hackathon bringing together experts in AI, computer vision, audio processing, and broadcast production to build a modular, real-time pipeline that detects key moments and automatically generates highlight videos. It combined speech-to-text (transcribing live commentary), audio classification (crowd reactions, music, silences), OCR (reading on-screen scores and timers), and motion/scene analysis (replays, celebrations, gameplay intensity). These inputs powered an automated editing engine producing social clips (30-90s), narrative recaps for streaming (3-5 min), and Smart Catch-Up features. Built on an open, scalable architecture with human feedback loops and compliance to standards such as EBUCore. Initially designed for football, it can adapt to tennis, Formula 1, e-sports, or archival video indexing.
6 people · senior-led, blended · onshore · 6 months · time_and_materials
public
Configure an engagement and see your own indicative band - free, no email required.
Amaris Consulting · Part of Mantu. Figures shown are indicative and illustrative ranges based on the inputs you provide, not a quote. Comparisons refer to typical market ranges for equivalent seniority, not to any specific firm.