Turning insights into action: How Sportsbet is transforming the betting experience

Sportsbet’s Ask AI lets customers pull up stats and insights so they can build their own bets using plain language.

Martha Sevigny, Sportsbet’s Senior Product Manager, explains how the tool provides bet suggestions and why control remains firmly – and deliberately – with the customer. 

The brand’s AI assistant was launched for racing bets in April 2025 and expansion to other sports began in December. It remains in beta, available to around 30% of customers, with plans to extend it to everyone. Sevigny explains what the team has learned from putting the assistant in front of real customers.

Ask AI Phone UX 2

1. What made Sportsbet decide to build this and are customers using it as you expected? 

Last year a working group was set up to explore how generative AI could solve real customer problems and pain points. One promising finding was that customers wanted to reach betting options quickly, with something behind the suggestion to explain it. The research kept returning to two words: quick and easy. So we  looked for an efficient way to build their bets while also building confidence in the selections through added context. 

For the most part, customers are using it as expected and their questions and feedback are shaping the roadmap. They ask a question in natural language and get back stats and insights, as well as suggested options they can take, edit or simply review. The curiosity is there, and so is the appetite to use AI in building a bet.  

2. Part of this is search and part bet construction. How do they fit together? 

I think of it as an AI-powered Bet Builder, backed by stats and insights drawn from the same sources as in the rest of the app. As a customer, I can ask about team or player performance for background research or ask the tool to suggest a bet on whatever parameters I give it. For example, if you’re betting on basketball in the US, that can be as specific as naming the legs – different parts of the bet – such as “Bronny James to score more than 20 points and the Lakers to win,” or as loose as “build me a five-leg bet for tonight’s Lakers game.” 

The tool is designed to be explicit about its limits. When it can’t answer a question, it will say so and offer related information where possible. There are also disclaimers in the product stating this is for entertainment, because we do not give advice. The simplest way I think about it is that the AI suggests, then the customer decides. It is completely in their hands whether to proceed or not. 

3. What does a customer actually see when the tool suggests a bet?

We built Ask AI to help customers explore sporting events and understand the factors that may sit behind a betting option. When a bet is suggested, customers see supporting information explaining why that option was generated. That may include relevant insights and, where available, recent performance for the team or player involved. Our aim is to give useful context so customers can make informed decisions.

Ask AI can suggest options, but the choice to place a bet, change it or walk away stays with the customer. Plenty of people use it purely to explore stats and insights. Responsible gaming messaging is built into the experience, and Ask AI sits within Sportsbet's existing responsible gaming framework.

4. How does responsible gaming factor into each interaction?

Customer safety is a key consideration in how we designed Ask AI. It includes safeguards intended to identify certain queries that may indicate a customer could benefit from additional support. When those safeguards are triggered, Ask AI does not provide betting suggestions, and the customer is directed to responsible gaming information and support resources.

These measures work alongside Sportsbet's broader responsible gaming tools, controls and customer protection processes. Ask AI is designed to support those protections, and responsible gaming considerations remain central to the experience.

5. What has putting this in front of live customers taught you that testing could not, and what did FanDuel contribute from their own AI bet builder? 

There is nothing like learning with a live product. We are still in beta, but real customer interactions have been invaluable and have driven how we refined the model and experience. People may say they want something, but you learn so much from how they actually choose to engage. 

The clearest example is the suggested prompts – pre-written queries so that customers don’t have to write their own. Examples include “Build me a three-leg NBA multi”, “Give me AFL tips for the weekend”, and “Which horse has the best record on wet tracks?” They account for around half of all queries, which makes getting them right a priority. Another example is that we started in racing, helping customers make sense of form quickly, and it was their feedback that then took us into other sports. 

FanDuel helped a great deal there, sharing research from their own in-app tool, Ace AI, so we could compare how customers used each product and work through shared problems such as building trust in AI. It was really collaborative and a strong example of the Flutter Edge in terms of how our brands’ insights can be leveraged across the group. 

6. Where does Ask AI go from here? 

I would like AI to help us build experiences that excite our customers and keep pace with how they naturally search, explore and bet. Beyond that, we want to keep building AI use cases that are of real use to them. The technology is moving quickly, so there will be a lot more we can do.