Accelerating content production with generative AI, featuring the NFL (AIM211)
Accelerating Content Production with Generative AI: The NFL's Journey
Overview
The video discusses the use of Amazon Lex (Q) in the NFL's media operations to streamline content production and knowledge management.
Key takeaways:
Amazon Lex (Q) is a generative AI-powered assistant that can answer questions, provide summaries, generate content, and complete tasks based on enterprise data and information.
The NFL faced challenges with onboarding and training a large seasonal workforce, as well as providing efficient access to complex technical workflows and documentation.
The NFL evaluated various solutions and ultimately chose to implement Amazon Lex (Q) Business to address these challenges.
The NFL's Problem Statement
The NFL has over 250 content creators and production staff, many of whom are hired seasonally.
Onboarding and training this workforce was a significant challenge, with 3-5 hours of training provided annually.
The organization has hundreds of complex workflows and use cases, and employees often struggled to find the right information to execute their tasks efficiently.
Delays in accessing information could lead to missed airtime and impact the content delivery process.
Implementing Amazon Lex (Q) Business
The NFL evaluated multiple generative AI solutions and chose Amazon Lex (Q) Business due to its ease of deployment, wide range of data connectors, and built-in security features.
The NFL went through a 3-month proof of concept (POC) process, which included:
Quickly deploying initial Q apps and iterating on them.
Refining and rewriting existing documentation to improve accuracy.
Organizing data sources into a structured taxonomy.
Testing different configurations and performing A/B testing.
By the end of the POC, the NFL's accuracy rate had improved from under 70% to around 93%.
Key Outcomes and Lessons Learned
The "Duke" Q app was rolled out to 300 users before the start of the NFL season.
It reduced the time to answer technical questions from 5 minutes to 10-20 seconds.
The organization was able to reduce onboarding training from 3-5 hours to just 1 hour.
Content creators could spend more time on creative tasks and less on technical minutiae.
Lessons learned:
Ease of deployment and the ability to start small and scale quickly are key benefits of Q.
Refining and restructuring data can significantly improve the accuracy and relevance of Q responses.
Investing in user training is crucial for successful adoption of Q within the organization.
Next Steps
The NFL is exploring additional use cases and stakeholders for the "Duke" Q app.
There are plans to integrate the Q app with the organization's Playbook Pro project to provide a unified user experience.
The NFL is excited about the new features and capabilities that will be introduced for Amazon Lex (Q) in the future.
These cookies are used to collect information about how you interact with this website and allow us to remember you. We use this information to improve and customize your browsing experience, as well as for analytics.
If you decline, your information won’t be tracked when you visit this website. A single cookie will be used in your browser to remember your preference.