How the Seattle Seahawks streamline game-day coverage using AI

As the Seattle Seahawks’ head coach steps away from the podium, the clock starts ticking. The post-game press conference, filled with insights on the team’s play immediately following a football game, needs to reach millions of fans across multiple platforms—and fast. In the past, this scenario would have set off a manual scramble in the Seahawks’ digital media department, with staff racing against time to transcribe, summarize, and distribute the content from the press conference across various channels.

But today, as those words still echo in the press room, a sophisticated generative AI-powered system is already at work, transforming the way the Seahawks handle their content distribution. This technological game changer, built on Amazon Web Services (AWS) and leveraging generative AI, is rewriting the playbook for how National Football League (NFL) teams manage their digital presence.

Kenton Olson, VP of Digital and Emerging Media for the Seahawks, is the architect behind this digital revolution. His team oversees all content published across the team’s online platforms, from their website and social media channels to their mobile application. They’re also responsible for managing these platforms, ensuring they’re primed for content creation and distribution.

The challenge Olson and his team faced was formidable. With over 1,000 videos produced annually, including more than 300 press conferences, the manual process of publishing content consumed valuable time and resources. Each video took an average of 30 minutes to publish, creating a significant bottleneck, especially on game days when multiple videos need rapid distribution.

As soon as the video team finalizes a piece—like head Seahawks’ coach Mike Macdonald’s post-game press conference—it’s automatically ingested into the system. AWS Elemental MediaConvert takes over, processing video to ensure it meets the team’s various format specifications. Processed video is then stored in an Amazon S3 bucket, ready for distribution across multiple channels.

But the system’s capabilities extend far beyond video processing. It automatically uploads content to the Seahawks’ website and YouTube channel, sends notifications, and even generates transcripts. The system also produces Spanish and German translations of content, expanding the team’s global reach.

The crown jewel of this new system, however, is its use of generative AI through Amazon Bedrock. This cutting-edge technology creates concise summaries of press conferences, complete with time codes for easy navigation. “Sometimes there’s someone who has not watched the press conference, and they’ll have to sit down and watch the entire press conference for half an hour, quickly turn around and write the copy, proofread it, and a number of other time-consuming tasks,” Olson says. “Now that’s all automated using generating AI technology from AWS.”

Figure 1: System architecture diagram with AWS components and services.

This automation is a game changer for the Seahawks’ digital team. It not only saves time but also allows staff to focus on more creative tasks rather than repetitive, manual processes. The system reduces a 60-step manual process to a streamlined, automated workflow that takes just a few minutes.

The benefits extend beyond just efficiency. The AI-generated summaries are more detailed and SEO-friendly than their human-written predecessors, potentially driving more traffic to the videos. The addition of time codes allows viewers to jump directly to specific topics within a press conference, improving and enhancing the user experience.

While the system’s current focus is on press conferences, Olson and his team are looking to expand its capabilities. Podcasts are next on the list, with plans to incorporate other long-form content pieces in the future. The potential applications seem limitless, with discussions underway about using AI to generate multi-channel social media campaign ideas.

One of the most exciting prospects is the potential to unlock the value of the Seahawks’ vast video archive. With the team’s 50th season in the NFL approaching, a wealth of historical content sits on servers and tapes, largely inaccessible due to the manual effort required to catalog it. The new system could process these old videos, making them searchable and bringing them back into the spotlight for fans to enjoy.

The implementation of this system wasn’t without its challenges. Optimizing the AI to understand team-specific nuances, such as unique name spellings, took time and effort. The team also had to build trust in the AI’s output, gradually moving towards a fully automated process.

The team improved transcription accuracy of player and coach names by implementing a custom vocabulary that updates in real time from the Seahawks’ roster. To address AI hallucinations, especially in longer press conferences, they provided the AI model with up-to-date lists of players and coaches, along with their positions and roles. The team also implemented an evaluation step to double-check for inaccuracies. To achieve more granular timestamps for easier navigation, they transformed the transcription output to include sentence-level timestamps. These solutions enhanced the accuracy and reliability of the system.

From a technical standpoint, the system is a testament to the power of cloud computing and AI working in tandem. It uses a suite of AWS services, including MediaConvert, Amazon S3, Amazon SNS, and Bedrock, seamlessly integrated to create a robust and efficient content pipeline.

While the system has already proved its worth, Olson is careful not to overstate its impact on fan engagement just yet. “I don’t know if it’s a good thing or a bad thing, but I think most fans probably don’t really recognize we’ve made a change,” he admits. However, he’s optimistic about the potential long-term benefits, particularly in terms of improved search engine referrals due to the richer metadata now associated with the team’s content.

The Seahawks’ innovative use of AWS and generative AI technology showcases the potential for AI to transform operations in professional sports. By automating time-consuming tasks and enhancing content distribution, the team is freeing up valuable resources to focus on creating engaging fan experiences.

As the NFL continues to evolve in the digital age, the Seahawks are positioning themselves at the forefront of technological innovation. Their AI-powered content distribution system is not just about efficiency; it’s about staying ahead in a competitive landscape where fan engagement is key.

The journey is far from over. As Olson and his team continue to refine and expand the system, they’re laying the groundwork for a future where AI plays an increasingly central role in sports media. From unlocking historical archives to potentially automating social media posts, the possibilities are as exciting as they are vast.

In the fast-paced world of NFL content, where every second counts and every fan interaction matters, the Seattle Seahawks and AWS are proving that in the digital age, the most exciting plays aren’t just happening on the field—they’re taking place in the cloud, where data, AI, and fan passion converge to create an entirely new ballgame.