Opens in a new tab
Southeast Asia Connect

News

AppsFlyer brings campaign data into marketers' AI workflows

Marketers can use AppsFlyer fraud, creative and integration data in their own AI assistants, while Agent Hub supports custom investigations.

· Lars Voedisch

Illustration of a Southeast Asian marketing analyst reviewing campaign data on a laptop beside conceptual AI workflow panels
Illustration: A Southeast Asian marketing analyst reviews campaign signals beside conceptual AI workflow panels.

AppsFlyer is opening fraud, creative-performance and partner-configuration data to marketers' own AI tools through MCP, a standard for connecting AI assistants to outside data and functions, while expanding Agent Hub so teams can build recurring analysis workflows. The 12 November release shifts more of the company's measurement work from a dashboard and reporting layer into customer-directed investigation. For app marketers comparing campaigns across Southeast Asian markets, the relevance is the ability to bring campaign evidence into the tools and routines their teams already use to make decisions.

The MCP release connects three kinds of evidence that usually answer different marketing questions: Protect360 fraud signals, creative-level performance cohorts and partner-integration settings. A team can examine whether suspicious traffic is concentrated in one geography, whether a creative retains or monetises users after installation, and whether attribution windows or postback rules explain a measurement discrepancy. Those questions can be asked through a marketer's own AI assistant using selected AppsFlyer data, rather than read only as separate dashboard outputs.

The expanded Agent Hub turns such questions into recurring work. A marketer can describe a performance workflow; AppsFlyer says an agent can analyse event-level data, investigate causes and send scheduled findings through Slack or email. The MCP tools extend selected data access to a customer's own AI client. Together, these changes let a team shape analysis around a problem it needs to revisit, such as a market-specific fraud spike or a creative that is losing effectiveness, and feed the findings into its operating process.

The release also sketches a narrower path from analysis to action. Its Apple Ads Share of Voice Agent can estimate visibility by keyword and geography, recommend a bid change and execute it only after the marketer approves. AppsFlyer says the bid-changing agent is in closed beta through a customer success manager. That controlled bid action is a limited example of the direction; the wider change is that measurement and campaign signals can now enter customer-built investigations and scheduled workflows.

Sources

AppsFlyer product announcement, 12 November 2026