Contrary to the popular narrative of shrinking teams, data suggests marketing headcount remains stable. The budget is being reallocated, with more spend going to people and a new, significant line item for AI tokens and inference costs, rather than being cut.
Marketing leaders should treat unpredictable AI token costs like a headcount expense rather than a fixed software cost. This mental model better reflects its variable nature, where spend is tied to work output, similar to an employee or contractor.
A new role, the "Super IC" or "zero-to-one marketer," will become central. This individual leverages a personal AI stack—comprising a context layer, connected tools, and custom agents—to execute a wide range of marketing tasks effectively, enabling rapid iteration.
A team of only AI-powered generalists will produce good but not great work. The optimal structure mirrors a renaissance workshop: "apprentice" generalists for rapid execution and "master" specialists who use AI to go deeper on high-impact, long-term strategic projects.
The traditional, hierarchical marketing org chart is obsolete because AI solves the context-sharing bottleneck it was designed for. Marketing should adopt the "pod" model from engineering: small, outcome-focused teams of marketers and AI agents working from a shared context layer.
Most teams assume they know where to focus. Instead, create a separate team of AI-powered generalists whose sole job is to cheaply test new channels and ideas. When they find a promising opportunity ("strike oil"), a specialist-led team is deployed to scale it.
