How One SMB Automates Content Localization with Generative Artificial Intelligence

Knowledge enriches our lives. There’s always something more that we can learn about the world around us, our cultures, and histories—even our futures. MagellanTV is a documentary streaming service committed to helping its viewers access stories that contribute to a lifetime of learning. Founded in 2018, this small business is based in Armonk, New York, where 20 talented employees manage a library of over 3,000 documentaries.

As MagellanTV Co-Founder Thomas Lucas explains, “Our core mission is to develop a relationship with documentary lovers on a global scale.” To do this, MagellanTV collaborated with Amazon Web Services and AWS Partner Network consultant, Mission Cloud, to turn language barriers into bridges with generative artificial intelligence (AI). The result is a powerful new AI-based workflow known as Magellan VoiceWorks.

Traditional methods of content localization have been a major barrier preventing MagellanTV from entering global markets. The business worked with a third-party service to help them translate and transcribe text for local markets, but high per-minute charges across thousands of hours of video made this unsustainable.

As Lucas explains, “Most of our programming is in English, and many of our viewers don’t speak English fluently. The traditional solution is dubbing, but that’s a time-consuming, costly, labor-intensive process. One language, one dub, can cost thousands of [US] dollars.”

All that changed when the company connected with AWS and Mission Cloud.

The value of working with cloud computing experts

Mission Cloud is a born-in-the-cloud managed services and consulting provider at the forefront of generative AI technology. It helps innovative companies scale existing generative AI models or build new, groundbreaking applications. Mission Cloud’s team of AWS experts empowers businesses to migrate, manage, modernize, and optimize their cloud environments, ensuring a successful future in the cloud.

Magellan TV initially took advantage of Mission Cloud’s free AWS security assessment, but deeper discussions uncovered an opportunity to help globalize content with generative AI. The first step was to conduct a comprehensive assessment of how AWS tools and services could be combined to automate video dubbing, translation, and transcription. Mission Cloud made use of the AWS Migration Acceleration Program (MAP) during this phase to accelerate results and control costs.

A collaborative adventure in generative AI

Considering the intricate nature of documentaries—layered narration, diverse voices, and distinctive pacing—it became clear that straight machine translation would not be enough to provide viewers with the experience MagellanTV wanted to give them. For example, a sentence might only be a few words in English but much longer when translated to German or Mandarin. There was also the question of cultural sensitivity—slang and idioms often translate poorly between languages, and phrases that might be perfectly innocent in one language may have negative connotations in another. MagellanTV needed a solution capable of handling all these elements efficiently at a scale involving thousands of documentaries and over a dozen languages.

Together, Mission Cloud and MagellanTV designed an automated pipeline for content localization powered by generative AI. This workflow combined multiple AWS services, including:

• Amazon Polly: Use high-quality, natural-sounding human voices in dozens of world languages
• Amazon Comprehend: Extract text, key phrases, topics, and sentiment for multiple uses
• Amazon Translate: Localize content with custom, high quality translations
• Amazon Transcribe: Automatically convert speech to text
• Amazon Bedrock: Generate text, summarize it, create preview images, and much more
• Amazon SageMaker: Build, train, and deploy machine learning (ML) models for any use case

According to Andrew Federowicz, Director of IT, MagellanTV, “Mission Cloud introduced us to its engineers, who proposed a plan to actually build this pipeline. Now fast forward to today. Vast improvements in generative AI systems, along with the evolution of AWS services such as Translate, Transcribe, and Polly, mean that it’s no longer ‘can we build and deploy this?’ It’s ‘how quickly’.”

MagellanTV is now busy building out this architecture, testing and ironing out any minor issues along the way, while already realizing massive time and cost savings. As Chelsea S. Lauzun, SVP of International Operations, explains: “We’re innovating. We’re building something that the world doesn’t have yet. We’re working with AWS to build something that will give our small company a truly global presence, a global reach.”

A lifetime of learning—localized

With a fully automated pipeline to translate, transcribe, and dub content, MagellanTV can vastly reduce the resources needed to share its content on a global scale. As Lucas explains: “This is a complex process. It relies heavily on generative AI. It has to replicate human translation and capture elements like slang, idioms, and culturally sensitive phrases. The Mission Cloud solution can do all this very quickly, within minutes, and that changes the economics of distribution.”

In addition to time savings, costs have also been dramatically reduced, “We were talking about somewhere like US$20 a minute to get content dubbed using traditional methods. Now we’re looking at less than US$20 an hour,” says Federowicz.

Lucas notes, “We come from the ranks of filmmakers, so our goal is to take people to new places, show it to them in a new way and make it accessible to them. It’s all about drawing together a global community of documentary lovers who share our curiosity about the world and our desire to explore, to understand, and see what’s out there.”

Conclusion

MagellanTV’s staff, like its viewers, are lifelong learners and their curiosity and willingness to explore the potential of generative AI has helped them build a solution capable of transforming their own industry and beyond. There are a few ways you can begin—or continue—your AI journey. First, learn about the most common use cases for SMBs curious about AI. If you need a third-party expert to help, seek out support from third-party experts. Still not sure and want to talk directly to AWS? Contact us to speak with a specialist.