SVG Sit-Down: AWS's Jason Dvorkin Talks Agentic AI, Cloud-Native Workflows, and the Future of Live Sports Production

SVG Sit-Down: AWS's Jason Dvorkin Talks Agentic AI, Cloud-Native Workflows, and the Future of Live Sports ProductionThe technology offers new workflows, ways of storytelling, broadcast efficienciesBy Jason Dachman, Editorial Director, U.S.

Thursday, September 25, 2025 - 9:46 am

Print This Story | Subscribe

Story Highlights

Few forces are shaping the future of live-sports production more than AI and the cloud, and Amazon Web Services (AWS) sits at the center of both.

While many organizations are still exploring how they can leverage generative-AI applications for live production, sports-media organizations are also turning their attention to agentic AI to streamline operations. From live cloud production to international-feed customization, these technologies are moving quickly from concepts to real-world impact - reshaping workflows, enabling new storytelling opportunities, and driving efficiency as broadcasters face rising rights costs and increased pressure to do more with less.

At IBC 2025, SVG sat down with Jason Dvorkin, AMER head of industry business development for media and entertainment, games, and sports, AWS, to discuss how agentic AI and cloud-native production are evolving, what use cases are gaining traction, and why these technologies are poised to redefine live sports coverage over the next 12-18 months.

What is the biggest trend or request you're seeing from your sports-media customers today?

We're seeing a lot of customers looking to understand more about agentic AI and where AI agents, which leverage foundation models to perform tasks to achieve predetermined goals, can help their operations. After lots of focus on generative AI for the past year, people seem to be drastically shifting to how they can implement agentic AI within their organization. What does that mean for content discovery? What does that mean for metadata application? And ultimately how does that help to better serve the fan?

What are you highlighting here at IBC?

We're showing a few different things with agentic AI, centered on our three main themes at the booth: creating content, connecting workflows, and captivating audiences. From a sports standpoint, we're seeing major growth in live cloud production for broadcast. That starts with getting content onboarded into the cloud and then looking at how it can be produced in the cloud. Agentic AI can come into play there to help with graphic localization for world-feed distribution or international broadcasting. We're also seeing a lot of work being done with agentic AI as it relates to storytelling and direct-to-consumer workflows.

Any real-world examples to offer on how agentic AI is being leveraged for live sports?

Sure. Our friends at the PGA TOUR have the entirety of their 27-PB archive on AWS and are now exploring how can they use agentic AI for storytelling across multiple platforms. They are using that video indexing and video understanding to pick out specific moments in content. For example, how can they use a plain-language search to find a clip of Viktor Hovland giving a ball to a fan or a specific moment on the 17th hole at TPC Sawgrass? An agent will find that specific asset, and another agent will use scoring data to turn that action into a story. Another agent can provide a notification that human review is needed to check the quality of the story. And that can get published instantly to the website or social media. On an event day, the digital teams can take more advantage of archive content or relevant storytelling from current day to package up something that can go directly to the fans.

With use cases like that in mind, have we passed the hype stage of Gen AI and agentic AI and on to the tangible benefit stage?

Absolutely. Here [at our IBC stand], over 70% of the demos that we have right now are using generative AI, and over 30% are using agentic AI in some fashion. We've seen how customers can use generative AI to create highlights to personalize that content based on their other data sets. But now it's starting to enhance that storytelling and offloading some of the more human-intensive manual work by using these agents to do some of that work so that humans can focus on more-meaningful tasks within their organization.

What do you expect in terms of the impact of agentic AI and generative AI on sports production and distribution over the next 12-18 months?

My belief is that we're not far from seeing agentic AI used for camera switching on a regular basis or for converting 16 9 video to 9 16 for platforms that require that. It's going to continue to evolve on a monthly basis, but you are definitely going to see some smaller sports-media organizations relying on AI rather than [deploying] a full crew to cover an event.

Can the AI be smart enough to say, Take this camera, roll this graphic, use this stats data, now go back to the game camera' and be able to tell the story? For high school sports or Division III College, I think that's going to happen very soon. You have the potential now to provide greater coverage and exposure to these sports but do it in a more cost-effective way for broadcasters and rightsholders.

You mentioned live cloud production as it relates to graphic localization or regionalization. How is that sector of the business evolving today?

If you were to try to translate graphics in a traditional broadcast workflow, you're talking about additional operators and expanding the production truck to support these additional resources. But now you can do that real-time translation using agentic AI tools and a generative AI component to create and render the graphic.

To reference PGA again, with their Every Shot Live [workflow], they can customize outputs for specific broadcasters. Now they can do the same thing for graphics as well. That way,

LINK:	https://www.sportsvideo.org/2025/09/25/svg-sit-down-awss-jason-dvorkin...
	See more stories from svg

Most recent headlines