How your organisation can leverage data and AI ‘for the win’

If you watch a Premier League football match broadcast today, you might be familiar with live statistical insights such as Win Probability – the chance that a team will win or draw a game.

Such match insights are enhancements that the Premier League, the top-tier football league in England and the most-watched competition in the world with an estimated audience of more than 3 billion people, have introduced to improve the broadcast experience for fans.

To derive win probabilities, 100,000 simulations of how a match will end are calculated every 30 seconds, using four years of match data.

These and other data insights are now possible after the Premier League started partnering with Oracle in 2021, said Aidan Mullally, Head of International Engagement for the Premier League, at a data and artificial intelligence forum organised by Oracle on 31 January.

As part of the collaboration, the Premier League, which is screened in more than 600 million homes in over 200 territories, and has broadcasting deals worth more than GBP10 billion (USD12.6 billion) between 2022 and 2025, has put 31 seasons’ worth of football footage – more than 1 million minutes of action – onto the Oracle Cloud.

“There is a lot happening in terms of how the Premier League uses AI to really assess what happens on the pitch and to predict outcomes,” Mullally said.

“And I keep coming back to a ‘personalised experience’ for fans. [We’re going to see] to what extent generative AI can really serve up what fans want before they even know they want it.”

Embarking on the AI adoption journey

“Regardless of where you are on this journey of AI adoption, I think what we all agree on is that the capability of AI has really accelerated over the last number of years,” said Chris Chelliah, Senior Vice President, Technology and Customer Strategy, Oracle Japan and Asia Pacific, at the forum.

“There is a tremendous amount of opportunity, there is a tremendous amount of potential. But it all starts with data. Data is the key ingredient in the AI recipe.”

For the Premier League, which now uses eight camera angles to capture footage in each of its 380 games per season, leveraging data has enabled it to introduce new player and team achievement awards, in addition to the match insights.

It now awards prizes for the Oracle Most Powerful Goal – the goal with the fastest average speed – and the Oracle Most Improbable Comeback, given to the team that comes back to win after having the lowest probability of victory.

“The questions, what we hear all the time [from customers], around AI are about how to scale AI to the enterprise level,” said Chelliah.

“There’s a perception that the capability, the skill sets are not there today, and the readiness isn't there is your organisation: ‘How do I adopt this thing, what are the risks, and how do we mitigate the risks?’”

Subscribe to the GovInsider Bulletin for the latest public sector and innovation updates.

Cutting through the confusion

C Prasanna Venkatesan, Director of Outbound Product Management, Generative AI Services, at Oracle said that many customers are confused by the many AI models available and coming onstream on the market.

“The theme and the use cases that we're working on with our customers have been on productivity savings. So how do I actually take my data, how do I maximise the value, and how do I make my work faster? And how can I generate some new insights, as an added bonus?”

Venkatesan highlighted the fact that the AI landscape is evolving so quickly that it’s no wonder customers are confused about implementing AI in their own organisations.

“Six months ago, we were talking about RAG or retrieval-augmented generation. Then we started talking about functions. Then we started talking about tools and now we're talking about agents and then it's going to be some other new term that's going to come up.

“So, there's a lot of confusion in the minds of our customers on how to actually implement this and put it into production.”

AI-embedded applications and finetuning foundational models

According to Venkatesan, Oracle has announced it is embedding generative AI into all of its software-as-a-service applications, with features to automate work processes such as writing recruitment job descriptions or writing emails.

“As a people manager, one of the most ‘hated’ things that I do now is writing all my annual performance reviews!” he said. “We are now actually automating it. I don’t know how authentic they will be, but we are doing it. We realise that it is actually a pain point for our customers as well.”

He added that Oracle is cognisant that not all customers would want to run all their applications or keep all of their data on Oracle.

“In order for our customers to still use some of the gen AI capabilities that we are bringing forth, we have the Oracle Cloud Infrastructure (OCI) Gen AI Service, which we just went into general availability earlier last week, and we also launched a beta service on Gen AI agents.”

Venkatesan added that customers could choose from several high-quality pre-built foundational AI models offered by Oracle, or they could customise or finetune a model, to better understand a specific industry or business context.

“You can think of [a pre-built model] as someone who's done their Bachelor's in English literature, so they're very good at answering questions. They can research anything that is a very generalist type of a question,” he said.

“When you finetune a model, what you're doing is that you're giving it a master's education in [your industry context, such as] logistics, or in video conferencing software.”

A simple AI formula for positive outcomes

“I think if you summarise what AI can do for an organisation, it is one of three things really,” said Chelliah.

“AI is really good at looking for similarities, it’s really good at looking for anomalies, and it's really good for being creative, getting some creativity into it. But now, what you want to do with all of that is to do it with the right governance.”

Chelliah said that Oracle can help customers build a boundary and a governance structure around the risks that come with utilising AI, such that organisations can maximise the use and value of their data.

“I always say AI is a simple formula: AI equals D plus I – AI is made up of data and infrastructure – and you need to bring the two together holistically to get [the right] outcomes.”