On September 28, CRM vendor Fast Track announced a new capability for its Fast Track AI: agent-to-agent. Previously, this AI could only work within its own interface, but now AI assistants in operators' other systems can directly request data and conclusions from it, without humans having to shuttle information back and forth.
Fast Track counts as a key system within operators, with hundreds of users running their CRM, player engagement, gamification, real-time intelligence, and daily decisions on it. The product lines Rewards, Greco, and True Value are built around an intelligence layer that helps operations see what is happening and where to act. After this update, that intelligence layer is no longer locked inside its own interface.
Anyone who has done API documentation integration knows: opening interfaces is easy, but explaining context clearly is hard. What Fast Track emphasizes this time is not exposing a few interfaces, but that its AI already understands its own ecosystem—which capabilities work together how, what the operational context looks like—and can work within the governance and workflows defined by the operator. When an agent from another system asks, "How is this player doing lately?" it directly provides a judgment, and the asking party does not need to understand Fast Track's data structures.
CEO Simon Lidzén put it plainly: organizations are moving toward a world where humans and agents work together, and expecting one agent to understand the internal workings of all systems is unrealistic. The way forward is for each to guard its own domain and be able to collaborate with one another. This release is also categorized under Fast Track's long-term roadmap of building an AI-native platform—not adding a chat window to the interface, but treating agent collaboration as a foundational platform capability.
Teams currently evaluating platform solutions can note this: in the future, connecting CRM, risk control, and data middle platforms may not necessarily rely on stacking up interface documents one by one, but could also involve their respective AI assistants talking directly to each other. During selection, it is worth asking one more question—can the vendor's intelligence layer be invoked externally, and how are governance boundaries drawn? To discuss the openness of various platforms in detail, you can contact us.