GGWP announced wider availability of its AI moderation and community safety platform for iGaming operators on 28 September 2026, following deployments with customers including Sportsbet and Tipico. The company announcement focuses on managing behavior across social features, with moderation decisions connected to operator policies and review processes.
For product teams adding chat or community features, the relevant development is the scope of that oversight. A player interaction can span several messages and formats before a moderator has enough context to assess it.
Behavior across messages and channels
GGWP says its platform combines content detection with user history and records linking an incident, the relevant policy and the resulting action. It also flags attempts to draw players into private channels, including disguised contact details. The company identifies scams and recruitment to unlicensed products among the risks that can follow those conversations.
Its moderation documentation describes several controls behind that approach. Usernames are screened when created or changed, while reputation persists through renames. Player reports are ranked using credibility and context, with player history attached, rather than being prioritised solely by the number of complaints.
The same documentation says live voice is transcribed and assessed using the text models. Higher-risk signals are escalated to people. These are descriptions of product capabilities, rather than published measurements of their accuracy in an operator environment.
Community moderation addresses a different set of signals from behavioural player-risk monitoring such as Mindway AI’s GameScanner. Where both are used, operators would need to establish who handles a conversation that also raises a player-protection concern.
What operators would need to test
Our assessment is that the practical test is whether the system gives a reviewer enough evidence to distinguish persistent abuse from an isolated misunderstanding. A warning, a temporary restriction and a human investigation carry different consequences. An operator should be able to examine why one was chosen and correct an inappropriate response.
Evaluation would therefore need to cover the languages, slang and communication formats used by the actual player community. Useful measures include false positives, missed incidents, time to human review and the outcome of challenged decisions. Teams also need to establish who owns an escalation when the conversation crosses a supplier-operated chat service and the operator’s own support channels.
GGWP’s announcement names existing deployments but provides no operator-specific evidence of improved retention or reduced harm. Those outcomes remain to be demonstrated separately from the availability of automated moderation.