Self-serve advertising is no longer merely automating media-buying tasks. AI is increasingly involved in which ads are delivered, whom they reach, what they cost, and which results appear in reports. For businesses, this raises a management question more important than tool selection: when platforms hold much of the optimization mechanism, where must the brand retain control?
On August 6, 2026, Gartner forecast that by 2028 more than 70% of global advertising spend and 80% of U.S. advertising spend will flow through self-serve advertising platforms in which AI materially influences media buying, cost, and outcomes. This is a forecast, not evidence of present performance for every business. It nevertheless reflects a real shift: the platforms’ back-office AI is becoming more deeply involved in audience selection, delivery, and pricing. That is distinct from generative AI used to write copy or create images.
Platform optimization is not the same as business optimization
Platforms have substantial advantages in behavioral data, testing speed, and bid adjustment. They can reduce manual work and find effective delivery combinations for the objective that has been set. Yet a system that performs well against an in-platform metric may still fail to produce a proportionate commercial result. A campaign may reduce cost per form completion while increasing irrelevant inquiries, for example; or it may lift sales recorded in a short attribution window without creating returning customers.
For that reason, “letting the algorithm learn” should not mean withdrawing from management. The platform should be allowed to optimize inside a defined boundary. The business must retain the commercial objective, the quality of input signals, brand guardrails, and the method used to judge genuine incremental value.
A control situation
A retailer asks a platform to minimize conversion cost. The system may concentrate delivery on people already familiar with the brand and more likely to buy. The conversion metric improves, yet the campaign may not create any new customers. If the commercial objective is customer-base expansion, the team needs to distinguish new from existing customers in measurement and use an appropriate experiment to determine whether spending generates incremental sales or merely reallocates sales that would have occurred anyway.
Three areas a business must own
The first is objective architecture. Translate the commercial objective into signals the platform can receive, but do not let an easily measured metric replace the intended outcome. If the business needs customers with high lifetime value, optimization signals should progressively reflect post-purchase quality rather than stop at the first purchase.
The second is independent measurement. Gartner recommends prioritizing platforms that support transparency and independent evaluation. This need not require an elaborate measurement system. A business can start with consistent source tagging, conversion data reconciled with sales, and a few geographically, temporally, or customer-group-based tests where scale permits. The aim is to obtain enough evidence to distinguish advertising-created outcomes from outcomes that might have happened anyway.
The third is platform-portfolio governance. Not every new platform deserves a large budget, but a business should not depend only on a familiar interface. It needs to identify the few platforms that truly matter to priority customers, invest management and creative capability there, and use other channels deliberately to extend reach, test an assumption, or preserve options.
What this means for marketers
As AI becomes more deeply involved in media buying, the marketing team’s competitive capability is not greater manual adjustment. It is the ability to ask the system the right questions: which objective should be optimized, which data is reliable, which result needs verification outside the platform, and which decisions must not be delegated entirely to an algorithm. This work connects marketing with finance, sales, data, and brand governance.
For small and medium-sized firms, the sensible next step is not to imitate the measurement systems of global companies. Select one or two important commercial objectives, standardize the input data, compare performance across customer groups or channels, and record what changes after each optimization cycle. This small but regular discipline builds better judgment than relying only on the platform’s default reporting.
Conclusion
Gartner’s forecast does not say that AI will solve advertising by itself. It indicates that optimization power is moving more deeply into self-serve platforms. In that setting, businesses need to retain control of objectives, data, independent measurement, and their channel portfolio. When those boundaries are clear, AI can make advertising more adaptive without weakening the brand’s decision capability.
Reference
Gartner. (2026, August 6). Gartner predicts more than 70% of global ad spend will flow through AI-influenced self-serve advertising platforms by 2028. https://gcom.pdo.aws.gartner.com/en/newsroom/press-releases/2026-08-06-gartner-predicts-more-than-70-percent-of-global-ad-spend-will-flow-through-ai-influenced-self-serve-advertising-platforms-by-2028
