An AI feature is production-ready only if the provider can show what data trained it, who monitors it, and how it was reviewed for your regulated markets. Anything less is a demo. Operators comparing platforms are asking about AI more often, and it now belongs in any igaming software development or vendor-selection discussion.
What is happening
Providers now describe AI across the stack: player-risk scoring, personalised offers, automated support, and trading models in sportsbooks. Some of these run live in production. Others are a scripted walkthrough on a staging environment. This piece covers the theme, not a specific announcement, so check current market activity yourself.
Why it matters
These features touch player protection, marketing compliance and customer data. A risk-scoring model that cannot be explained is a problem when a regulator or auditor asks why a player was or was not flagged. A personalisation engine that pushes offers at vulnerable players is a liability, however good the demo looked.
What to do about it
Put these questions to every provider in writing:
- Training data: What data was the model trained on? Is it trained on your players’ data, pooled data, or neither?
- Live or roadmap: Can you see it running in a production environment with a current operator, or only in a demo?
- Oversight: Who monitors drift and false positives? Can a human override a decision, and is the override logged?
- Regulatory review: Has the feature been reviewed against the rules in your licensed jurisdictions, particularly for responsible gambling and marketing?
- Explainability: Can the system show why it scored a player or chose an offer?
- Ownership: If AI is part of a managed service, who is accountable for its decisions, you or the provider?
- Exit: Can you switch a feature off without breaking the platform?
Run the same checks on sportsbook features. Ask whether odds and trading models are in-house, licensed from a third party, or still planned.
Sourcecodelab is a B2B iGaming platform provider, so we are one of the vendors these questions apply to. Ask us the same things you ask everyone else, and read our answers as a vendor’s, not an independent review. Compare them against what you can verify yourself: documentation, references and a live environment.
