Everyone’s Talking About AI Agents Running Your Marketing Department. Almost Nobody’s Doing It.

Almost Nobody’s Doing It.

There’s a version of AI marketing you keep seeing here on LinkedIn: autonomous agents planning campaigns, launching them, and optimizing themselves while the CMO sips coffee and watches a dashboard. It makes for a great post. It is not, for the vast majority of us, what’s actually happening inside real marketing organizations right now.

There is that story and the data… and a gap we are not talking about. I think the gap itself is the more useful story, and because I’ve spent a bit of time now building the unglamorous stuff that actually moves pipeline. Guess what… it wasn’t an autonomous agent. It was infrastructure.

The numbers say we’re mostly still assisting, not delegating

BCG surveyed 300 global CMOs this year and found that almost all of them, 96 percent, say AI is driving end-to-end transformation of their function – that is not surprising. But when BCG looked at what teams had actually built, only about a third had done the real work. Most marketing AI use today is still humans directing a tool on a single task. Less than a third of teams have moved into agent-led workflows, and only 8 percent are running campaigns where multiple agents operate on their own.

That’s not a knock on marketing leaders. It’s a pretty honest snapshot of where the technology and the org design actually are.

“Agentic with a small a” is the more accurate headline

Digiday ran a piece earlier this year built around a phrase I can’t stop thinking about: agentic with a small a. The idea is that most of the AI wins CMOs are actually banking right now are not autonomous decision-makers, they’re efficiency plays. Campaign taxonomies. Ad tracking. Naming conventions. Resizing creative. Automating the repetitive trafficking work that used to eat a junior team member’s whole week.

None of that is flashy. All of it is real. And the industry sources quoted in that piece made a point worth sitting with: layering more automation on top of already opaque systems, especially in places like programmatic media buying, can make oversight harder, not easier, if you don’t fix the underlying structural problems first. More AI is not automatically a fix for a broken process. Sometimes it’s a faster way to run the broken process.

Where I think the real opportunity actually is

Google’s Think with Google published a piece this summer from a Kellogg marketing professor Jim Lecinski making a point I’d push CMOs to sit with before they redesign their org charts: the mistake most leaders make is assuming AI will simply make the team smaller. The more useful question is what shape the team needs to become. Execution-heavy roles shrink. But you need more people who can orchestrate hybrid workflows, supervise agents, set standards, and act as the standard-bearers for judgment and taste that a model can’t replicate.

That matches exactly what I’ve seen rebuilding a marketing function post-merger this year. The AI didn’t replace strategy. It replaced the parts of the job that were never really strategy to begin with, and it freed up real hours for the parts that are.

Three questions I’m asking before I call anything “agentic”

1. Is a human still making the decision, or just reviewing one the system already made? Those are different claims, and CMOs should be precise about which one they’re making publicly.
2. Is the underlying data foundation actually unified, or is the agent making calls on fragmented inputs? Because a confident answer built on bad data is still a bad answer.
3. Would I be comfortable explaining this workflow to my CFO in one sentence? If it takes a paragraph, it’s probably not ready to run unsupervised.

The CMOs who are going to look smart in two years are not the ones who claimed the most agentic transformation the fastest. They’re the ones who were honest about the distance between the pitch deck and the pipeline, and closed it methodically.

Where are you actually on that spectrum, assisting or delegating? I’d genuinely like to know what “agentic” looks like on your team right now.

Sources: BCG, “Moving the Agentic Marketing Transformation from Illusion to Reality” (2026); Digiday, “‘Agentic with a small a’: CMOs are adopting AI more slowly than it’s evolving” (Feb 2026); Think with Google, “How CMOs should rewire their AI marketing team,” Jim Lecinski (2026).

Originally published on LinkedIn on July 13, 2026 → Read the original on LinkedIn