Agentic conversational video intelligence built on AWS
Learn how to build a conversational video intelligence solution on AWS using an agentic architecture. A single Strands Agents SDK agent orchestrates Amazon Bedrock, Amazon Rekognition, and Amazon Transcribe at runtime, deciding which service to call so you can ask natural language questions about your videos and get answers in seconds.
In an innovative approach to video intelligence, AWS enables organizations to ask natural language questions about uploaded videos and receive answers in seconds. With the growing volume of video content in various sectors like media, security, insurance, and professional services, traditional methods of manually reviewing hours of footage have become impractical. The solution presented here utilizes an agentic architecture that dynamically determines which AWS services to invoke based on the user's natural language query.
Instead of creating custom machine learning (ML) pipelines for each specific question type, which would require significant development effort, the system employs the Strands Agents SDK to orchestrate Amazon Bedrock, Amazon Rekognition, and Amazon Transcribe. This single AI agent can handle a variety of use cases, such as extracting design decisions from meeting recordings, identifying individuals in security footage, or summarizing portions of videos.
The agent makes runtime decisions on which services to use, invoking Amazon Transcribe for spoken content, Amazon Rekognition for face matching, and reusing cached results for follow-up questions about previously processed content.
The implementation is available on a companion GitHub repository, and a major media and entertainment company has successfully adopted this approach, reducing manual review time by approximately 80 percent for over 200 multi-hour recordings. To implement this solution, users need an AWS account with access to Amazon Bedrock (with Anthropic Claude Sonnet enabled), Amazon S3 for storage, and the necessary IAM permissions.
Python 3.11 or later with the Strands Agents SDK installed is also required. The agent orchestrator, built using the Strands Agents SDK and powered by Amazon Bedrock, receives natural language queries from users, determines the appropriate tools to invoke, sequences multiple service calls when needed, and synthesizes the results into coherent answers.
Written by urgent.news from AWS Machine Learning's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
