Agentic AI & the Boundaries of Automation

The most philosophically charged conversation in AdTech this week centered on agentic advertising — and specifically, where human oversight must remain non-negotiable. An AdExchanger op-ed argued that the industry must proactively define what AI can never touch before autonomous ad-buying systems become entrenched, raising questions about brand safety, creative approval, and audience targeting decisions that carry ethical weight. This isn't abstract: as agentic systems begin executing media buys, creative iterations, and audience segmentation with minimal human input, the absence of clear guardrails creates legal and reputational exposure for advertisers. The piece arrives at a moment when competitive pressure from platforms like Salesforce's Agentforce for Marketing is intensifying, with virtually every major marketing automation player — Braze, Klaviyo, HubSpot, Adobe Journey Optimizer, and others — now positioned in direct competition with AI-native marketing agents. The industry is being forced to answer a governance question it has largely deferred.