Published: 2025-06-02 - Authors: Jack Bradham and Meera Balasubramaniam.
Organizations that manage large audio and video archives often struggle to extract value from their media libraries. Imagine a radio network with thousands of broadcast hours across multiple channels–they must verify ad placement, identify interview segments, and analyze programming patterns. In this post, we show how you can automatically convert unstructured media files into searchable, analyzable content. By combining Amazon Transcribe, Amazon Bedrock, Amazon QuickSight, and Amazon Q, organizations can achieve the following goals:
In the sections that follow, you will see how these AWS services work together so your organization can tap into the full potential of its media, whether you need to ensure advertising compliance, analyze content, or locate specific moments within thousands of hours of recordings.
This solution provides an event-driven media analytics pipeline that transforms how you manage and capture value from your content:
The following diagram illustrates the architecture.
This event-driven architecture automatically processes and analyzes media using AWS services. The workflow includes the following steps:
This architecture transforms raw media into searchable, analyzable data while maintaining an organized structure for efficient access. The event-driven design ensures that new uploads are handled automatically, and the combination of AWS AI services unlocks a deeper understanding of the content. Each AWS service contributes to this transformation:
While this post focuses on the technical implementation of the media analytics pipeline, every production deployment must include comprehensive security controls.
Secure storage in Amazon S3
Identity and access management
Network security
Data encryption
Monitoring and detection
Access control
Compliance requirements and governance policies may affect how you implement this solution. Consult AWS security best practices and partner with your security team to tailor the controls to your use case. For more guidance, see Best Practices for Security, Identity, & Compliance.
To follow along, you need:
This solution uses three distinct buckets to support the media analytics workflow:
Follow the Creating a general purpose bucket guide to provision each bucket.
Enable event notifications on the raw input bucket so EventBridge can trigger automation. The system watches for activity in the S3 buckets–when new content is uploaded or transcripts are available, EventBridge invokes the right Step Functions workflow. For details, see Creating rules that react to events in Amazon EventBridge.
Example filter for newly uploaded media:
{
"source": ["aws.s3"],
"detail-type": ["Object Created"],
"detail": {
"bucket": {
"name": ["rawinputbucket"]
}
}
}
Example filter for newly added transcripts in the data lake:
{
"source": ["aws.s3"],
"detail-type": ["Object Created"],
"detail": {
"bucket": {
"name": ["business-data-lake"]
},
"object": {
"key": [{
"suffix": ".transcription"
}]
}
}
}
The orchestration layer includes two primary workflows. The first handles media ingestion and transcription, while the second manages AI analysis. Each workflow includes safeguards for potential errors and retry logic. For a walkthrough, see Learn how to get started with Step Functions. One diagram illustrates how new media is ingested, indexed, and transcribed, and another diagram shows how transcripts are analyzed downstream.
Launch Amazon Transcribe jobs with the permissions you configured. Speech-to-text features such as language detection, speaker identification, and custom vocabularies improve accuracy for your media. Refer to How Amazon Transcribe works for setup details.
Drive the AI analyzer by crafting precise prompts that extract meaningful information. Amazon Bedrock examines transcripts to pinpoint segments, speakers, and topics, turning raw text into structured data. For guidance, see Design a prompt.
Example prompt:
You will be reviewing a radio transcription to identify advertisements and extract relevant details. Your task is to analyze the provided transcript and output the results in a specific JSON format based on a given schema.
Please follow these steps to complete the task:
1. Carefully read through the entire transcript.
2. Identify all advertisements within the transcript. Look for clear indicators such as product mentions, promotional language, or transitions from regular content to commercial content.
3. For each advertisement you identify, determine the following information:
- Company: The name of the company being advertised
- Start time: The timestamp in the transcript where the ad begins
- End time: The timestamp in the transcript where the ad ends
- Product: The specific product or service being advertised
4. Format your findings into a JSON object that follows the provided schema. Each advertisement should be a separate object within an array.
5. Ensure these fields in your response are provided for each advertisement.. All are required fields: company, starttime, endtime, product.
6. Use precise timestamps for start and end times. If exact times are not available, make a best estimate based on the transcript's context.
7. If a particular field is unclear or not explicitly mentioned in the transcript, you may use "Unknown" as the value.
8. Only respond with json and nothing else. Do not provide comments or explain your answer.
9. Surround the JSON response with standard ```json markers
Here's an example of how your output should be formatted:
{
"advertisements": [
{
"company": "TechGadgets Inc.",
"starttime": "00:05:30",
"endtime": "00:06:15",
"product": "SmartHome Hub"
},
{
"company": "FreshFoods Market",
"starttime": "00:15:45",
"endtime": "00:16:30",
"product": "Organic Produce Delivery Service"
}
]
}
Do not add any fields that are not specified in the schema, and ensure all required fields are present for each advertisement.
Adopt a hierarchical data-organization strategy to enable efficient storage and analytics. Use AWS Glue crawlers to discover and catalog media metadata automatically. For more details, see Using crawlers to populate the Data Catalog.
Configure Athena tables so you can run SQL-based queries on the media insights. Example view:
CREATE OR REPLACE VIEW "commercials_view" AS
SELECT
metadata.market market,
metadata.station_call station_call,
metadata.format_type format_type,
CAST(metadata.timestamp AS timestamp) timestamp,
ads.company adCompany,
ads.product adProduct,
ads.starttime,
ads.endtime
FROM
(commercials
CROSS JOIN UNNEST(advertisements) t (ads));
Enable natural-language interaction with the media archive using Amazon Q Business. Configure knowledge bases and metadata so users can search and access content through conversational queries. Use the processed insights bucket to populate the knowledge base. For setup, see Getting started with Amazon Q Business. The referenced screenshot shows example conversations with an AI assistant.
With QuickSight, create visual analyses that make your insights tangible. Connect to Athena views to highlight ad trends, content analysis, and performance metrics through interactive dashboards. See Tutorial: Create an Amazon QuickSight dashboard for a complete walkthrough.
After you deploy the architecture, execute these activities to ensure the solution meets your performance and business goals.
Establish comprehensive testing
For example, select a recently processed radio show and compare its transcript with the original audio. Review the AI-generated insights to confirm that major events–such as ad transitions or interview segments–were detected correctly. To ensure the system handles all content types, sample a diverse mix, including morning talk shows, evening news, and weekend sports programming.
Benchmark performance
Monitor how long it takes to process different file lengths–from quick ads to full-length shows. Observe resource-consumption patterns to pinpoint bottlenecks, such as slower transcription for certain formats or opportunities to optimize parallelization.
Validate real-world experience
Have team members interact with Amazon Q using the questions they would normally ask, like “interviews about climate change last week.” Collect feedback from distinct personas–content managers versus compliance reviewers–to refine the system.
This structured testing approach, combined with real-world scenarios, lays the groundwork for a robust, user-friendly media analytics solution.
As you transition from initial deployment to production, optimization becomes critical for both cost control and user satisfaction. A network that processes thousands of hours per week can unlock major savings and discovery improvements by refining transcription accuracy or throughput. Marketing teams analyzing ad placement also depend on precise insights. Consider the following optimization strategies:
Start with a pilot deployment and expand gradually as you mature your media-analytics capabilities.
To avoid incurring ongoing charges, remove the resources you created after testing:
Multiple industries can leverage this architecture to mine value from audio and video content. Tailor the solution for your needs, such as broadcast management, corporate communications, educational archives, and more.
By harnessing AI, organizations can turn raw audio and video into structured insights that enable faster, better-informed decisions.
This article demonstrates how to use AWS services to convert unstructured media into actionable intelligence. By combining Amazon Transcribe, Amazon Bedrock, QuickSight, and Amazon Q, you can build an automated, scalable media-analytics solution tailored to your organization’s needs. The architecture provides:
Organizations can now transform content into searchable knowledge, extract insights automatically, develop data-driven content strategies, and streamline operations through automation. As audio and video libraries continue to grow, the ability to process and extract value efficiently becomes increasingly vital. This architecture delivers a solid foundation for today’s requirements while remaining adaptable to future innovations.
We invite you to explore how media analytics with AWS AI services can address your organization’s unique challenges. Identify your top use cases and unlock the insights hidden across your media archives.
Jack Bradham is a Senior Solutions Architect at AWS with more than 20 years of leadership experience in technology. Before joining AWS, he held key roles at Microsoft and Google, advising federal agencies and Global 100 enterprises. Jack holds an MBA in International Business from the University of South Carolina and has deep expertise in cloud computing, enterprise architecture, and business transformation. He is passionate about helping customers design scalable cloud solutions that meet business goals through innovation.
Meera Balasubramaniam is a Senior Solutions Architect and Data Analytics Specialist at AWS. With over 20 years in technology, she helps organizations turn complex business challenges into actionable data solutions. Meera specializes in scalable data architectures, advanced analytics platforms, and cloud solutions. She brings extensive experience in enterprise data strategy, business intelligence, and machine learning, speaks frequently at industry events, and mentors the next generation of cloud experts.