campaign ideas and assets, providing the team with a steady stream of new content to test. The second wave added further intelligence and safeguards, with agents running rapid pretests of creative concepts and automatically checking content brand, legal, and risk compliance. The final wave extended the system globally, enabling agents to adapt messages for local markets and coordinate scalable testing and rollout. Together, these waves transformed a slow and manual process into a fast and data-drive system that, in some content creation pilots, increased the speed of the end-to-end process by four times versus traditional workflows. Agentic systems are also beginning to emerge in media execution. One advanced advertising platform is now building AI agents to autonomously optimize campaigns across major digital channels, continuously evaluating performance, adjusting bids and budgets, pairing creative with audiences, and generating new message variants. These agents operate in real time, managing thousands of microadjustments that previously required constant manual oversight. Early adopters report faster optimization cycles and measurable improvements in return on ad spend, highlighting how agentic execution is reshaping modern media operations. Fueling growth and adoption, while limiting risk End-to-end agentic workflows will help marketing organizations capture value by producing more consumer experiences far more quickly, while powering top-line growth and fueling working spend. But facilitating this change is no simple task, requiring leaders to execute in key ways across the organization. Brands will need to set a top-down vision (led by the board and CEO), with strong governance to ensure adoption and scaling, while limiting brand and legal issues. Leaders also must understand that agents are only one tool in the AI playbook; other tools, including scripting, robotic process automation, and machine learning, also need to be considered. Focusing too narrowly on agents alone can leave significant efficiency gains on the table when scaling. Nor is this process without risk—especially in marketing, which directly affects consumer-facing content and brand perception. Marketers will need to pay close attention to potential brand and legal vulnerabilities, above and beyond the technology and data risks posed by agentic AI across all functions. Marketers seem to understand the novel risks AI presents. A McKinsey survey of 35 CMOs of Fortune 250 consumer and technology companies found that executives were primarily concerned about brand and legal governance, human capability challenges, technology under investment, and data bottlenecks. Insights teams will also need new governance mechanisms to validate AI-generated insights, establish confidence thresholds, and ensure accuracy before findings inform major brand or investment decisions. Nearly 90 percent of CMOs are experimenting with AI use cases across various points of the marketing process, but less than 10 percent have captured value across end-to-end workflows, McKinsey research has found. Agents can help move the needle. But as they begin to deploy Reinventing marketing workflows with agentic AI 9