The Productivity Paradox of Agentic Advertising

AI Agents Meet the Solow Paradox

Forrester argues that the promise of “agentic advertising” collides with the classic Solow paradox: massive tech spend does not automatically translate into higher labor productivity. The blog notes that in 2000 researchers found a $1 hardware investment often required up to $9 in process redesign, training and hierarchy changes.

Redefining Core Advertising Workflows

The post outlines three workflow shifts. First, media planning will evolve from a static, periodic document to a “continuous decisioning” system that ingests campaign history, audience intelligence, media costs, creative test results and business outcomes. Planners will move from building plans to setting the rules that trigger plan adjustments. Second, media buying will transition from platform-centric interfaces to intent-driven instructions, where traders specify objectives, constraints and brand-safety parameters for agents to execute. Third, media optimization will be anchored to business outcomes—lifetime value, margin, incrementality—rather than traditional KPIs, demanding high-quality proprietary signals to gain a competitive edge.

Preparing Teams for Machine-Mediated Execution

Forrester recommends three immediate actions. Reorient planning around discrete decisions—such as reallocating spend between channels—and codify the thresholds that trigger agent recommendations. Build an “agent-ready” architecture that makes first-party audiences, conversion signals and creative metadata machine-readable across demand-side platforms, retail media networks and open-web publishers. Finally, reskill practitioners to interrogate agent moves, differentiate genuine performance signals from attribution noise, and override agents when short-term ROAS conflicts with longer-term incrementality.

The blog concludes that firms that wait for agents to mature before redesigning workflows risk falling behind competitors already leveraging continuous decisioning, machine-speed buying and always-on optimization.