Marketing ops · Production

5 ways Prime Video improves the viewing experience with generative AI on AWS

The problem

Prime Video's growing content library made it harder for customers to find what they wanted, while marketing assets were scattered across disparate systems with insufficient metadata, impeding content discovery, rights tracking, quality control, and monetization.

Workflow diagram · grounded in source
1
Personalized content recommendations
ai_action
“Amazon Bedrock is helping power personalized content recommendations directly within the "Movies" and "TV Shows" landing pages of Prime Video. Viewers will see "movies we think you'll like" and "TV shows we think you'll like" collections…”
2
X-Ray Recaps video analysis
ai_action
“X-Ray Recaps works by analyzing various video segments. Combined with subtitles or dialogue, it generates detailed descriptions of key events, places, times, and conversations”
3
Spoiler-free guardrail check
validation
“Amazon Bedrock Guardrails are applied to verify that summaries remain spoiler-free”
4
Prime Insights sports analysis
ai_action
“Prime Insights illustrate hidden aspects of the game and forecast pivotal moments before they happen. For instance, the "Defensive Vulnerability" feature on TNF is powered by a proprietary machine learning model. It employs thousands of …”
5
Rapid Recap live event catchup
ai_action
“Rapid Recap automatically compiles a full recap of highlights, up to two minutes in length, then drops fans into the live feed”
6
Dialogue Boost audio enhancement
ai_action
“Dialogue Boost analyzes the original audio in a movie or series and uses AI to intelligently identify points where dialogue may be hard to hear above background music and effects. Then, the feature isolates speech patterns and enhances a…”
7
Media asset metadata extraction
ai_action
“Media2Cloud on AWS guidance—which delivers comprehensive media analysis at frame, shot, scene, and audio levels. The guidance helps to enrich assets' metadata (such as celebrity, text on screen, moderation, mood detection, and transcript…”
8
Metadata sync to MAM
integration
“Metadata from media assets of Prime Video are automatically fed to an AWS Partner, Iconik, media asset management (MAM)”
Reported outcome

Prime Video enriched hundreds of thousands of assets and improved discoverability in its marketing archive, while generative AI on AWS now powers personalized recommendations, spoiler-free recaps, sports insights, audio accessibility, and video understanding across the platform.

Reported metrics
Media assets enrichedhundreds of thousands of assets
Marketing archive discoverabilityimproved discoverability in its marketing archive
Dialogue Boost additional languages supportedsix languages
Reported stack
Amazon BedrockAmazon SageMakerAmazon Bedrock GuardrailsAWS BatchAmazon ECRAmazon ECSAWS FargateAmazon S3Amazon DynamoDBAmazon CloudWatchMedia2CloudAmazon NovaAmazon RekognitionAmazon TranscribeIconik
Source
https://aws.amazon.com/blogs/media/5-ways-prime-video-improves-the-viewing-experience-with-generative-ai-on-aws/
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Prime Video enriched hundreds of thousands of assets and improved discoverability in its marketing archive, while generative AI on AWS now powers personalized recommendations, spoiler-free recaps, sports insights, aud…

What tools did this team use?

Amazon Bedrock, Amazon SageMaker, Amazon Bedrock Guardrails, AWS Batch, Amazon ECR, Amazon ECS, AWS Fargate, Amazon S3, Amazon DynamoDB, Amazon CloudWatch.

What results were reported?

Media assets enriched: hundreds of thousands of assets; Marketing archive discoverability: improved discoverability in its marketing archive; Dialogue Boost additional languages supported: six languages (source-reported, not independently verified).

How is this marketing ops AI workflow structured?

Personalized content recommendations → X-Ray Recaps video analysis → Spoiler-free guardrail check → Prime Insights sports analysis → Rapid Recap live event catchup → Dialogue Boost audio enhancement → Media asset metadata extraction → Metadata sync to MAM.

Related marketing ops cases