skip to Main Content

How AI Supports Content Management in Rally

Content management is a critical function for media supply chain platforms. Data comes into the media supply chain in many forms, and finding the resources to enrich and modify it can be a difficult and time-consuming task for an Operations team.

However, as artificial intelligence has made notable leaps in quality and ability over the past few years, becoming incredibly accurate and far more reliable, it can now be relied upon as a tested, trusted way to enrich and access content data. By leveraging natural language processing (NLP) and machine learning (ML) techniques, AI can identify trends, predict user needs, and offer recommendations, enriching both the management of content and the user experience.

Many SDVI Rally workflows already onboard AI:

  • Transcription
  • Language identification and verification
  • Automatic metadata enrichment
  • Fine tuning based on requirements (ex: creating summaries of different length, extraction of actors, identification of genres, etc.)
  • Identifying localization data (Language analysis, celebrities, logos, brands, violence, nudity, etc.)
  • QC data (Scene changes, black frames, etc.)

These workflows ensure that operators can focus their time on exceptions found by the AI tools, rather than searching for problem areas. In this way, AI tools streamline processing of content and help operators focus on the work that is of highest value: human tasks.

Examples with AWS Text Rekognition

AI-generated bounding boxes

Using the time-based metadata markers generated by AWS Text Rekognition, we can highlight areas of the screen where text occurs. This is useful to:

  • Determine if text is positioned correctly
  • Determine whether burned-in text has spelling errors
  • Determine whether text encroaches on out-of-bound areas
  • Navigate on screen to see the timecode where the text occurred (example: finding logo or burned-in subtitles)

AWS Transcribe

AWS Transcribe gives you the ability to visualize what the actors have said and to quickly jump to key dialogue or an important moment. Having this information at your fingertips helps you make creative, edit, or QC decisions faster. Or, take it a step further: export the Transcribe markers into a JSON file to be used in an upstream process, such as converting them to subtitles!

If you’d like to find out more about how AI tools can increase efficiency and improve your Rally media supply chains, reach out to your TAM or solutions architect.

Back To Top