Mon–Fri: 9:00 AM – 6:00 PM (UK)
All case studies
Thomson Reuters
News & Information

AI-powered video summarisation at scale

Built an automated pipeline that ingests broadcast video, transcribes, and produces editorially-formatted summaries — cutting analyst review time dramatically.

The challenge

Editorial teams were spending hours manually reviewing long-form video content to extract key moments, quotes and summaries for downstream products.

What we built

1

Designed a serverless ingestion pipeline on AWS to handle large video files with chunked, fault-tolerant processing

2

Combined Amazon Transcribe for STT with foundation models on Bedrock for editorial-style summarisation against a custom prompt-and-style guide

3

Added a human review surface so editors could accept, edit, or reject AI output and feed corrections back into the pipeline

Outcomes

Analyst time per video dropped from hours to minutes

Pipeline scales to peak news cycles without intervention

Editorial voice preserved through prompt engineering + reviewer feedback loop

Stack
AWS Bedrock
Amazon Transcribe
AWS Lambda
Step Functions
S3
Ready when you are

Ready to Automate with AI?

Schedule a free consultation with our team of ex-AWS specialists. We'll discuss your challenges and outline how AI-powered automation can transform your workflows—no commitment required.