Kicking goals with AI for the AFL
As a front runner in diversifying and innovating across multiple touchpoints, technologies and audiences, the Australian Football League faced a hurdle that we at Truis were more than excited to tackle. In the midst of a global pandemic, our task was to develop an AI solution that would help the AFL continue to meet sponsorship agreements and launch a platform of new opportunities for current and future sponsors.
Image supplied by AFL Photos
“We needed to make AFL On Demand work! Truis played a huge part in making it happen.”
When COVID-19 pulled the curtain on live entertainment, the AFL had to act fast to develop new ways to entertain footy fans and meet sponsorship agreements. The idea was to create an entertainment platform that provided unlimited access to an extensive catalogue of AFL content series, documentaries, short films and historical footage. They called it, AFL On Demand – the magic of footy, all in one destination. But as they say, with every new idea comes new challenges, this was certainly the case for the AFL’s broadcasting and media teams when piecing together their new entertainment hub.
Hugh, the AFL’s lead Broadcasting Engineer, came to us with a problem. AFL On Demand enables fans to access an extensive catalogue of new and old AFL content. Any content with ‘match vision’ or ‘gameplay’ had to display the AFL and Toyota joint logo on screen to meet sponsorship agreements before launching the online platform. This meant digging through thousands of minutes of footage and overlaying the logo at varying moments. The AFL faced a manual, time-heavy process that would have to be repeated if sponsors changed in the coming years. With deadlines approaching and an almost impossible task ahead, a budget-conscious Hugh looked to AI to get the project over the line.
“Choosing Truis for the job was simple. Straight away, they understood what we needed and proved they could solve the problem – we just knew that we were going to get what we were asking for.”
From the start, we were thinking about what this project meant for the AFL and their agreements with Toyota – It was essential to make it an easy payback for them. The first step was defining the problem and developing an MVP or proof of concept to share with the AFL. Our team built an AI model, ran existing AFL footage through the model, and inserted the joint logo where necessary – providing this early concept allowed the AFL to gain complete confidence before undergoing any further steps.
Once all parties were on board, we gathered the necessary data and processed over 30 thousand videos using our in-house machines – this ensured all the data was safe and didn’t leave our control. We applied a time-mark with an AI machine learning algorithm, so when ‘match vision’ or ‘gameplay’ started, the AFL and Toyota joint logo would appear. The time-stamp was seamlessly integrated into the AFL’s video matrix system. When subscribers watch, the video matrix grabs the time-stamp in real-time and places the logo on the screen. Using technology such as Kubernetes and AC922 enabled a fast, and efficient solution to the AFL’s challenge.
“We came to them with a problem, and Truis came back with a solution that worked, and they made the whole thing easy.”
Through a collaborative process, we could minimise risks and build a flexible solution specifically for the problem. Thanks to a tailored built AI solution, the AFL’s On Demand platform met deadlines, avoided hours of manual work, and met all sponsorship agreements. The AFL can now easily update sponsor logos and deliver even more on-screen time for their partners in years to come.
Many businesses out there don’t understand the potential of leveraging their data to gain business potential or solve a problem. The AFL saw an opportunity to use AI and avoided an almost impossible manual alternative.
“Working with Truis was more personable than most IT engagements I’ve had in the past with external companies – everything was clear, it was laid back, it was friendly, and we always felt understood.”