By Jonathan Baker Amazon is upgrading its application of AI in a big way. The iconic e-commerce brand is allowing the always-advancing technology to oversee campaign execution at a greater scale. This is a recent example of an emerging trend which may be called “agentic AI delegation.” Efficiency Requires Oversight Amazon’s new Ads Agent shows how quickly AI is moving from assistant to decision maker. While the gains in efficiency may inspire awe and appear attractive, marketers need governance frameworks and human oversight to ensure those systems remain compliant, transparent, and aligned with business goals. Unboxing Amazon’s New Plan Amazon Ads, the platform where businesses and sellers promote select brands and products across Amazon as well as other external sites, is upscaling its agentic advertising practices. This plan was announced at “unBoxed 2026,” an annual event that teaches advertisers how to make the most of their presence on the platform. This is what attending advertisers learned… Presenting Amazon Ads Agent Advertisers who once relied on Amazon DSP and Ads Console will now conduct operations through the new Amazon Ads Agent. They will interact and ask questions through the Agent via conversational chat that simulates natural language. The Agent can also assist with campaign planning, execution, and optimization, while still allowing marketers to maintain various levels of control. And while on the topic of campaigns… New Campaign Categories Amazon Ads Agent provides three specific campaign categories to help advertisers accurately organize their product offerings: “Sponsored Ads,” “Full-Funnel Campaigns,” and “Display, Video, and Audio (DVA+).” (Side note regarding “Full-Funnel Campaigns”: Early open-beta results show 67 percent higher long-term ROAS and 29 percent lower new-to-brand acquisition costs.) Impressive? Yes. And yet Amazon, despite its fame and recognition, is but one of many organizations that are adopting greater agentic integration. The efficiency gains are impressive, but they also raise questions about oversight, transparency, and accountability. From Assistant to Operator For several years since its inception, AI capabilities were limited to pilot and evaluation phases. More recently, many businesses have “promoted” AI to full integration responsibilities that include automating advertising workflows. According to findings from EMARKETER, 43 percent of marketers in the United States are following this trend, with 50 percent of agencies applying agentic AI to campaign execution. While many may marvel over the above percentages, it is essential for advertisers and marketers to stop and take necessary precautions before they decide to go “all agentic AI” by practicing the following: Using Guardrails to Guide The need to place guardrails at the center of future AI-controlled advertising campaign plans and executions has never been greater. Asset-level review remains important, but marketers should increasingly focus on system-level guardrails that oversee how autonomous AI operates, makes decisions, and stays within compliance boundaries. Taking guardrails to the next level… Double Down on Disclaimers Advertisers and marketers must take extra care ensuring that disclaimers accompany any AI-created or altered content. (Laws mandating this are already in effect in New York and were recently passed in California. It is likely only a matter of time before these laws are applied to the rest of the nation.) A Final Thought The “human factor” is the foundation of the two precautions described above. To place this importance in perspective, according to the World Economic Forum, fewer than 1 percent of organizations have fully operationalized responsible AI practices. Amazon and other well-known companies will hopefully serve as an example to the rest of the world and apply agentic AI practices responsibly. Do you have questions about how to balance AI implementation with proper caution? EGC Group can help you. Fill out the “Let’s Talk” form, and someone will be in touch with you shortly.
