AI
AppsFlyer introduces AI agent for Apple Ads bidding and expands marketing automation tools
AppsFlyer has expanded its AI marketing tools with customizable agents, new data integrations and an Apple Ads agent that can adjust campaign bids after human approval.
(For more news, click here)
AI is becoming more widely used in digital advertising, helping businesses manage campaigns and make faster decisions using data. However, companies still need to understand AI recommendations and stay in control of decisions that affect their marketing budgets.
AppsFlyer announced on November 12 an expansion of its agentic AI capabilities, introducing customizable marketing agents, four new Model Context Protocol (MCP) capabilities and its first AI agent designed to take action directly within an advertising platform.
The updates extend AppsFlyer's Agent Hub, allowing marketing teams to automate performance analysis, investigate campaign problems and access marketing data through AI tools. A new Apple Ads agent can also recommend and apply bid adjustments, although changes require approval from a marketer before execution.
AI agents move from analysis to campaign execution
The expanded AppsFlyer Agent Hub allows marketers to create customized AI agents by describing the tasks they want performed in natural language.
These agents can analyze campaign and event-level data, investigate performance changes and deliver findings through Slack or email according to a predefined schedule. Teams can also review the agent's proposed analysis plan before deployment and examine the reasoning behind its recommendations.
For example, a marketer could configure an agent to investigate declining conversions, identify possible causes and recommend actions based on campaign performance.
The release also introduces the Apple Ads Share of Voice Agent, AppsFlyer's first agent capable of implementing changes rather than only generating recommendations.
The agent monitors advertising visibility across selected keywords and geographic markets, estimates share of voice and recommends bid adjustments when performance falls outside an advertiser's target range. Once approved, the changes can be applied directly to Apple Ads from AppsFlyer.
The feature is currently in closed beta and available through AppsFlyer's customer success managers. According to the company's documentation, it supports campaigns with manual keyword bidding, while automated bidding campaigns are not supported.
"Marketing teams are moving from manual campaign orchestration to agentic execution," said Inna Weiner, VP Product at AppsFlyer. "With this release, AppsFlyer is giving marketers trusted, agentic tools that can analyze, recommend and act on their data to deliver clear value for growth."
New MCP capabilities expand access to marketing data
AppsFlyer has also introduced four MCP capabilities that allow marketers to work with more of the company's platform data through compatible AI assistants and external tools. MCP is an open standard that connects AI applications with external data sources and services, allowing users to retrieve information and perform supported tasks through conversational interfaces.
The first addition provides access to Protect360 fraud detection data, helping teams examine suspicious advertising activity, identify geographic concentrations of fraud and investigate which validation rules blocked potentially fraudulent traffic.
A second capability provides creative-level performance information, including advertising expenditure, reach, conversions, retention and revenue. This allows marketers to compare creative assets, identify declining performance and determine which campaigns may benefit from optimization.
Two further tools focus on campaign configuration and troubleshooting. OneLink Link Creation allows marketers to generate campaign tracking links from natural-language descriptions, with approval required before creation. Partner Integration Settings provides access to attribution windows, event mappings and postback configurations, helping teams investigate integration issues without manually navigating multiple settings pages.
The capabilities extend AppsFlyer's existing MCP offering, which connects marketing data with external AI applications. Some functions, including creative performance access and deep-link creation, were already introduced in earlier product updates. AppsFlyer's MCP server remains documented as a beta service, despite individual capabilities being listed as generally available.
Growing demand for AI-driven marketing workflows
The expansion comes as more marketing teams experiment with autonomous workflows. An AppsFlyer report on AI usage found that the number of companies using its MCP infrastructure increased twelvefold during the first five months of 2026. The report examined activity from 1,914 companies connected to AppsFlyer MCP and 1,543 companies that configured AI agents during that period.
The findings suggest increasing interest in automating recurring marketing analysis, although they reflect activity within AppsFlyer's own customer base rather than the wider industry.
In the Middle East, AppsFlyer sees opportunities for agentic AI across gaming, fintech and e-commerce, where businesses frequently manage campaigns across multiple advertising channels and markets.
However, introducing AI into campaign execution also raises questions about access permissions, data handling and the reliability of automated recommendations. Human approval provides an additional safeguard, but does not eliminate the need to validate AI-generated analysis before making spending decisions.
AppsFlyer's latest updates reflect a broader transition in marketing technology, with AI tools increasingly moving beyond reporting toward operational tasks. Their effectiveness will depend on whether businesses can automate routine decisions while retaining sufficient visibility and control over campaign performance and expenditure.
