Religious Institution Case Study

‍

Overview: Protecting Irreplaceable Media for Future Generations

Some media archives contain both operational value and historically irreplaceable assets. For a historic religious institution in Manhattan, New York, that reality has shaped every decision about how to store, manage, and surface its massive collection of audio, video, and still image assets.

The Client is one of Manhattan’s most storied institutions, with deep roots in American history. Its grounds and community have been associated with notable figures and significant events spanning centuries. Today, the Church remains an active community presence, hosting public events, supporting civic causes, and welcoming visitors from political, religious, and public life.  

Managing millions of assets (including one-of-a-kind recordings of historic moments) with legacy storage tools was no longer sustainable. The Client partnered with Nomad Media, an AWS Partner, to modernize their media asset management platform and bring the power of generative AI to their archive.

The Challenge: Preserving and Surfacing Decades of Irreplaceable Content

For years, the Church’s media library lived across a patchwork of storage solutions: local hard drives, external backup drives, thumb drives, and a modest cloud footprint. It was a practical approach for its time, but it created logistical problems as the archive grew. Assets were difficult to locate, at risk of being lost to hardware failure, and essentially invisible to the staff who needed them most.

The stakes here are different from a typical media organization. Many of these recordings have never been re-broadcast or re-released. They document historic events, community milestones, and the work of a living institution with centuries of presence in its city. Losing them, or even being unable to find them, becomes both an operational inconvenience and a loss of institutional memory.

‍

Archivists and promotional event producers needed a way to centralize media asset storage, improve media discoverability across a sprawling library, and find specific content without manually reviewing hours of footage.  

The Solution: Cloud-Native Media Management, Upgraded with Generative AI

The engagement began with the adoption of the Nomad Media Platform as the Client’s master cloud asset management system. This was a foundational shift that centralized the library and brought it into a modern, scalable environment powered by Amazon Web Services. With that foundation in place, the next step was clear: apply AI metadata generation to make the archive searchable.

As an AWS Partner, Nomad Media brought both platform expertise and hands-on AWS integration knowledge to the deployment. The team worked closely with the Client to assess the expected improvements in search capability, plan the rollout, and ensure AWS service costs were well understood before going live.

The Retrieval-Augmented Generation (RAG) feature was deployed through an upgrade to the Client’s existing Nomad Media installation. Activation is controlled by a customer-accessible admin switch, giving the Client full authority over when the system enters production mode, ensuring no disruption to ongoing workflows.

Because the RAG system operates automatically in the background, end-user training was light. Nomad Media trained a designated admin using a train-the-trainer approach, keeping the rollout lean while ensuring the team was ready to take full advantage of the platform’s enhanced capabilities.

The deployment was completed on a short timeline. From go-live, archivists and event producers gained a fundamentally different relationship with their media library, one where intelligent video search, AI-generated metadata, and semantic understanding replaced manual browsing and guesswork.

Technical Overview: AI-Powered Media Archive on AWS

How the AI Metadata Generation Pipeline Works

The Nomad Media Platform’s Retrieval-Augmented Generation system is built to understand content, not just label it. Rather than relying on manually entered tags or filenames, the RAG pipeline automatically discovers rich, semantically aware metadata for every ingested asset. This means staff can search by meaning, context, and content rather than hunting through folders or relying only on finding what someone labeled it years ago.

Once media is ingested and transcoded by the Nomad Platform, the following AI-powered operations run automatically on each asset:

• Transcription and subtitling powered by Amazon Bedrock Data Automation

• Scene segmentation: dividing media into discrete video and text segments with AI-generated scene summaries, using Amazon Bedrock Data Automation

• Image-to-text summarization via Amazon Nova

• Vector and embedding generation from text content, using Amazon Nova

• Vector and embedding generation from image content, using Amazon Nova

• Storage of all generated metadata in Amazon DynamoDB, indexed for Amazon OpenSearch

Custom AWS Lambda functions orchestrate each stage of the pipeline, with API calls triggering processing steps and AWS SQS handling messaging between the Nomad Media Platform and AWS services. The entire workflow runs without manual intervention once the RAG feature is active.

Amazon Bedrock and Amazon Nova also scale automatically with processing demand, ensuring consistent performance whether the system is processing a single asset or working through a large backlog without inflating infrastructure costs.

AWS Cloud Infrastructure: Services & Integration

The RAG pipeline is built on three core AWS services: Amazon Bedrock, Amazon Nova, and Amazon DynamoDB. These sit within a broader cloud-native media asset management platform that natively integrates Amazon Virtual Private Cloud, Amazon S3, Amazon Cognito, AWS Lambda, Amazon SQS, and AWS Storage Gateway.

For an archive of this scale and historical significance, the budget of long-term storage is important. Amazon Glacier reduces the total cost of ownership of the asset library, keeping irreplaceable content safely preserved without the cost of keeping everything on hot storage. Live content delivery supporting the Client’s ongoing public events and community programming is handled by Amazon Elemental MediaLive, a broadcast-grade live video processing and packaging service, with Amazon CloudFront and Amazon CloudTrail providing world-class public distribution service. Amazon Bedrock Data Automation is also available to generate captions and enable per-user content localization for streaming and video-on-demand delivery.

Intelligent search across the archive is powered by Amazon OpenSearch, which indexes the AI-generated metadata from the RAG pipeline and makes every enriched asset immediately discoverable across the platform. Amazon QuickSight is also available within the platform as a choice for analytics and reporting.

Platform Integrations

The Nomad Media Platform’s RAG functionality is built entirely on AWS with no third-party AI or metadata services involved in that pipeline. For standard platform functionality, the Nomad Media application integrates Google AdSense and the Bitmovin HTML5 Video Player.

Staff access the platform through Nomad Media's SSO integration, simplifying user management and role assignment across the organization. The Bitmovin player supports internal media preview, allowing staff to review proxy copies of archived assets directly within the platform without requiring downloads or additional tools.

Results: A Living Archive That’s Actually Searchable

With Nomad Media’s AI-powered media management in place, the Client’s decades of collected media, including historic recordings, community events, and public programming became accessible in a way they hadn’t been before. The combination of AI metadata generation and RAG-powered search meant that finding the right asset no longer depends on institutional memory or manual review.

Archivists can now search by content and context, not just filenames. Promotional and events teams can surface relevant historical footage quickly to support new programming and community outreach. And the organization as a whole has a media asset management platform that reflects the care and permanence its archive deserves.

For an institution whose media library is also a historical record, the ability to preserve, discover, and thoughtfully deploy those assets is a meaningful part of honoring history.

‍

Related Reading

MAM and Live Streaming for Religious Organizations