The question I'm hearing most from marketing leaders right now isn't about tools.

It's about org design.

How do you restructure a GTM team when AI is changing the tech stack, required skills, and so many things related to the jobs to be done? What do you hire for versus what you reskill for? What does "AI-native" actually mean?

I've had versions of this conversation with a dozen leaders over the past month. Different companies, different stages, different pressures. But the disorientation underneath is consistent: they feel the scale of the shift. They're making real decisions inside it. But the playbook doesn't exist yet — and we are all learning and adapting in real time.

Here's the framework I'm building from, starting with the content function and working outward.

Every GTM restructuring conversation is really one question: when AI can handle more and more of the execution layer, where should human judgment live — and how do you organize around it?

That's not a technology question. It's an architecture question.

And the teams navigating it best aren't starting from "what tools should we implement." They're starting from: what does this function need to produce? Who makes the calls that determine quality? How does work flow from idea to impact? Then they build the AI layer around that. Tool selection is the last decision, not the first.

I spend my time inside the content function specifically, so that's where I can be precise. But content is a useful model for the broader GTM org — because the same forces are acting on it, just more visibly.

Before AI: a content team was made up of senior strategists and junior content writers. The bottleneck was production capacity.

Now: production capacity isn't as much of a bottleneck. A small content team with a solid workflow can ship what used to require larger teams and/or content agencies. The bottleneck moved.

Now the bottleneck is with judgment. Who decides what to write, for whom, with what point of view, measured against what outcome? Who reviews whether the output sounds on-brand, differentiated, worth a buyer's time? Who owns the feedback loop that makes the system sharper over time?

These aren't production questions. They're strategist questions. And they require the kind of contextual, stakeholder judgment that sits with the person who owns the outcome — not a model that owns nothing.

If production is less of a bottleneck, you're no longer hiring primarily for production capacity.

What you're actually building — in every GTM function, not just content — is the judgment layer. The senior operator who can brief the AI model with enough strategic specificity that the output doesn't need to be rebuilt. The person who can review against brand standards not with a checklist, but with the pattern recognition that comes from having done it a hundred times. The function owner who understands how to trace activity to pipeline and defend the function when budgets tighten.

In practice: smaller teams, higher individual leverage. Fewer people running more sophisticated workflows. The org chart gets leaner; the capability requirement per person goes up.

And the role of AI in the workflow isn't "replace the junior hire." It's "give the senior operator more execution capacity." Senior judgment, AI execution, a workflow that compounds. That's the architecture.

Nobody has fully cracked the skills and structure question for an AI-native GTM org. Not yet. What I see in the leaders who are furthest along is that they started with one function, built the system, documented what worked, and used it as a model for the next one. Not a top-down transformation. A working prototype that compounded.

The content engine is often the right place to start — fast to instrument, results are visible quickly, and the judgment-layer architecture it requires is the same one that scales across demand gen, product marketing, and the full GTM motion.

If you're navigating the restructuring question and want to think through what the content function looks like as a starting point: reply "STRUCTURE" and let's have the conversation.

— Jill

P.S. Next issue: the specific system I use to encode brand voice into Claude — the reference document structure, the senior review pass, and how the two compound over time. Useful if your AI content doesn’t sounds like you. [Ships Tue June 16]