Transforming sports betting: AI and the future | AGB
Summary
Sportradar’s Chief Product, Technology and AI Officer Behshad Behzadi explains how AI is no longer a marketing buzzword but a foundational component of modern sports betting and sports technology. The company uses computer vision, machine learning and generative AI to scale data collection, detect integrity issues, generate content and personalise user experiences.
Behzadi — a long-time Google veteran and co‑founder of Google Assistant and Google Lens — describes Sportradar’s mix of in‑house and third‑party models, including a detailed foundational basketball model able to simulate player behaviour at high temporal and spatial resolution. The tech is being applied across odds modelling, coaching and scouting, content production and targeted advertising, while governance and product limits guard against misuse.
Key Points
- AI is essential to Sportradar’s operations — without it they could not scale predictive analytics or odds creation.
- Computer vision (introduced in 2023) enables around 100× more data capture than human observation, powering richer models and integrity checks.
- Sportradar combines bought and home‑grown models, augmenting external foundations with proprietary sports data for better specialisation.
- They built a foundational basketball model using high‑frequency tracking (29 body points, 60Hz) to simulate player actions and compute granular metrics like expected possession value.
- Generative AI powers Content Studio to automatically produce articles, video and audio variants in real time for different audiences and languages.
- Bet Concierge is a chatbot that explains odds and offers smart bet suggestions — improving transparency and user trust.
- AI governance is crucial: Sportradar draws boundaries (for example around automated bonus-setting) to avoid harming clients or violating business logic.
- Personalised UI and emotion‑aware ad timing can retain younger users while protecting VIP bettors; targeted messaging improves engagement and conversion.
- Applications go beyond betting: coaching, scouting and remote analysis benefit from simulation and aggregated tracking data.
- Despite technical capability to offer thousands of bet options, customer demand centres on a limited set of popular markets — product design must reflect that reality.
Content summary
Behshad Behzadi frames AI as a practical necessity for firms processing vast sports data. Sportradar uses computer vision and machine learning to detect fraud, enhance odds accuracy and drive new products. Their generative tools create scaleable media output, while chatbots like Bet Concierge improve clarity around pricing.
The firm’s basketball foundational model demonstrates how granular tracking data can be converted into realistic player simulations that serve betting, coaching and scouting alike. However, Sportradar stresses governance and selective application — not everything should be automated, and letting models control business policies (bonuses, discounts) can be dangerous.
Finally, AI is leveraged for UI personalisation and time‑sensitive advertising by reading fan sentiment and optimising messages for moments when audiences are most receptive.
Why should I read this?
Short version: if you work in betting, sports tech or media, this is a neat snapshot of how a major data provider actually uses AI — not the hype. It explains where AI helps (odds, integrity, content, coaching) and where it shouldn’t run wild (bonuses, opaque decisions). Quick, practical and industry‑relevant.
Author style
Punchy: this piece matters. It shows AI being applied end‑to‑end in sports betting by a company with deep data assets and an engineering mindset. If you need to understand where investment and risk are headed in the next 2–5 years, read the detail.
