How AI Is Used in Sports Betting: Odds, Assistants, Streams and Integrity

AI in sports betting is no longer a pitch-deck idea. The products below are in use or on sale today, each named in a public document. For every one we say what it does, who is on record using it, and whether anything beyond the vendor’s word backs the claims. The full list, with sources, is in our AI products tracker.

1. Pricing markets

Odds have long been set with statistical models; the change is how much of the book is now compiled without a human trader. Kambi says its AI-powered trading division, Tzeract, has powered football pricing for the Kambi network since 2022 and supplies a Bet Builder product, with European operator kwiff named as its first commercial customer in 2024. In July 2026 Kambi said the World Cup was the first tournament at which all its pre-match and live offers were “fully compiled and traded” by its algorithmic system. Those are Kambi’s statements; it has not published an independent test of the prices.

2. Betting assistants

The most visible use is a chat box. FanDuel’s AceAI lets customers ask for player and team stats and build parlays in natural language; the answers come back as structured bets that can be added to the slip. Flutter says AceAI is live for around half of FanDuel’s US customers, was built in-house by a team of about 20 in roughly a year on large language models provided through Amazon Web Services, and that when a prompt suggests a problem, such as asking how to win back losses, it points to help and stops showing bets.

Suppliers sell the same idea to operators who cannot build their own. Symphony Solutions’ BetHarmony takes text or voice requests to place bets, check odds and answer account questions. Symphony claims it can cut support workload by 50–70%, but it names no operator using it.

3. In-play streams and sports data

Genius Sports’ GeniusIQ is the data and AI layer under its betting and media products: it ingests tracking and play-by-play data, which Genius says is captured up to 200 times per second, and turns it into contextualised data and video. One product built on it is BetVision, a single-screen stream that combines low-latency video, live stats and a bet slip with AI-generated overlays; Caesars Sportsbook and Fanatics Sportsbook were named as the first operators at its 2023 launch. Our BetVision explainer covers where the AI actually sits.

On the content side, Sportradar’s GenAI Automated Content API turns its sports data into AI-written match previews and recaps, published two days before a match and within 30 minutes of the final whistle, according to Sportradar.

4. Personalisation and offers

This is where most operator AI spend goes, and where the rules are tightest. DraftKings told investors in its 2025 annual report that “across our product offerings, we actively use data science and machine learning to help optimize conversion and monetization.” Suppliers such as Future Anthem sell this as a product: Amplifier AI tailors content, bonuses and interventions to individual players, with Betsson, GiG and Bede Gaming named as partners. CRM vendors now add natural-language agents on top, such as Optimove’s OptiGenie, Fast Track AI and Xtremepush’s XpertOS.

What other listed operators told the SEC about AI is on each company’s earnings page, for example Flutter and DraftKings.

5. Integrity and fraud

Sportradar’s UFDS AI scans odds from more than 600 bookmakers, plus account-level betting data, to flag suspicious patterns for sports federations and leagues. Sportradar says AI detected 977 suspicious matches in 2023, 73% of the total it found that year. Alerts are reviewed by analysts before anything is reported. We cover this and the anti-fraud tools that operators run in AI fraud and integrity detection in gambling.

6. Player protection

The same behavioural models that drive personalisation can flag players at risk of harm. FanDuel’s AceAI is one example. Specialist suppliers such as Mindway AI, Neccton and Playtech’s BetBuddy are compared in AI responsible-gambling tools compared.

What the rules say

• EU: Since 2 February 2025, the AI Act has banned AI that exploits vulnerabilities due to age, disability or a specific social or economic situation to distort behaviour in ways likely to cause significant harm. Since 2 August 2026, people must be told when they are talking to a chatbot.
• Massachusetts: Under 205 CMR 257.02, sports wagering operators may not use patron data to promote specific wagers or offers through algorithms, machine learning or AI that they know, or reasonably expect, make the platform more addictive.
• US federal: The SAFE Bet Act would bar operators from using AI to track an individual’s wagers, create offers targeted at specific individuals, or create products such as microbets. It was introduced in March 2025, and GovTrack shows no action since.

Every measure, with its status and next date, is in our AI regulation tracker.