Enterprise transformation

– 28 min read

The AI-native marketing org: Six functions that need to get done

Diego Lomanto, CMO  |  July 23, 2026

In the last year, I’ve watched marketing roles change faster than at any point in my career. Not in theory: in practice, on my own team at WRITER and on the teams of the companies we work with.

A content manager who used to write 10 blog posts a month now curates the system that writes 50. A campaign manager who used to build one campaign at a time now orchestrates five in parallel, directing agents instead of drafting copy. A brand strategist who used to be in every approval meeting now encodes the brand standards once and lets agents reproduce them across every output. The org chart and titles remain the same, but the work and the rewards shift radically.

Most of their organizations haven’t noticed. The org chart looks the same. The performance reviews measure the same things. The budget still flows to whoever produces the most volume.

That’s the problem I want to talk about. Not which AI tool to buy. Not whether agents are ready. Because they are. The question is whether your marketing org is structured and honest about what the highest-leverage work actually is now.

And to be clear upfront: this isn’t a prescription to reorganize your team. Not yet. What I’m going to lay out is a framework. Six functions every marketing org needs to perform in an AI-native world. The org chart can stay. People can keep their titles. What changes is what they spend their time doing, and what the organization rewards them for.

This isn’t a reorg. It’s a rebalancing of who does what.

The first thing to understand is that companies aren’t mature enough to reorg around this yet. And the data backs that up: 75% of executives say their company’s AI strategy is more for show than for actual internal guidance. If you don’t have a real strategy, you’re not ready to restructure around it. What’s happening instead is a rebalancing of who does what: agents take the execution, while people handle the deciding, the encoding, and the building.

From activity to outcome

It’s about humans owning the outcome, not just the activity. A social media manager who used to own the activity of writing and scheduling posts now owns the outcome: a brand presence across platforms that sounds like us, performs, and adapts as the platforms change. Agents took over the activity. The outcome stayed with the person: that’s the rebalancing.

The reward system hasn’t caught up

And that shift has to show up in what the org rewards. Most marketing orgs today promote execution volume, like how much got produced, how many campaigns shipped, and how many assets delivered. That’s what gets budgeted. That’s what gets celebrated. But when agents handle execution, execution volume stops being the measure of a person’s contribution. What matters now is the work that makes everything agents produce better. Curation and architecture used to be treated as overhead, squeezed in between the “real” work of shipping campaigns. Now they become the real work. The producing increasingly gets done by agents. The deciding, the encoding, the building: that’s where the leverage compounds.

A loop, not a project

And it’s not a one-time setup. Curation, architecture, and orchestration feed each other. The work builds on itself, the way a flywheel does: not in a single burst, but with every turn. That’s what makes it different from the old model, where execution was a series of discrete tasks with a start and an end. This is a loop.

How do you measure leverage?

Which raises the question every CMO asks me: how do you measure it? Output is easy to count. They’re posts published, campaigns shipped, and assets delivered. But curation is harder. How do you put a number on the person who made every agent’s output better? We’re working on that right now, inside our own org. We don’t have the answer yet. But we know the wrong answer is the one most orgs default to: measuring the activity instead of the outcome. A curator’s value isn’t in how many guardrails they wrote. It’s in whether the work the system produces sounds like your brand or generic noise. We’ll share what we learn as we figure it out.

Why the traditional marketing team structure breaks, and quick fixes don’t help

Most marketing teams today are structured around execution. Specialists own execution. Managers oversee execution. The hierarchy exists to ensure execution happens at quality and scale.

That was the right design for a world where humans were the execution engine. And it served us well. But when AI enters the picture, the broad base of people doing repeatable, high-volume execution suddenly has less to do. Agents can handle that work now. Whether they handle it well depends on something we’ll get to shortly: the context layer. For now, the point is that the work itself is no longer the constraint.

Cutting people is the wrong fix

When agents start doing the work your specialists used to do, the obvious move is to cut headcount. Some companies will, and in some cases, they should. But most leaders cut based on what disappeared (the execution) instead of what remains (the work of encoding organizational intelligence). Cut the people who shape your brand, and your system won’t go generic overnight; it will drift gradually because no one is keeping it sharp. Reshape their roles around that work instead.

AI doesn’t replace the skills these people spent their careers building: judgment, taste, strategic instincts, and brand fluency. It amplifies them. The social media specialist who built their craft around understanding what resonates on each platform now applies that craft at the system level, like curating what good looks like, reviewing where agents drift, steering the strategy, instead of spending it all on posts they personally write. Same craft, wider reach. Less time on the work, more time on what makes the work good.

Adding roles is only the start

I recently sat in a room with CMOs from some of the most recognizable brands in the world. Every single one of them told me the same thing: they know the way their marketing org works has to change. And most of them have done something about it. They’ve hired AI-specific roles. An AI marketing specialist. An AI innovation lead. A small team that owns the AI agenda.

That’s a start. Those roles matter. But adding a role doesn’t change how the other hundred people work. It lets the org say it has an AI strategy while everyone else keeps doing their job the way they always have. The data backs this up: 70% of CMOs say becoming an AI leader is a critical goal, but only 30% report the organizational readiness to actually scale it. The gap between aspiration and readiness comes down to how the org is designed, not just who it hires.

The hard part is reshaping the roles that are already there, like the specialists, the managers, the people whose jobs are already changing whether the org acknowledges it or not.

The AI-native marketing org: Six functions that need to get done

Today, marketing teams need to account for six distinct jobs-to-be-done. But the way those jobs get done, and who does them, is fundamentally shifting. Four of these jobs are human: strategy, architecture, curation, orchestration. Two are owned by agents: execution and real-time intelligence. The human jobs still need to get done. Some people will do all of them. Most will do two or three. Roles will be hybrid, and that’s perfectly fine. This is the AI-native marketing org, and it rests on six pillars.

Strategy

Setting the high-level direction. Defining the brand’s core narrative. Making the high-stakes judgment calls.

This is where the CMO and VP-level leaders live. AI doesn’t replace this. If anything, AI makes it more important because every agent deployed will operate within the parameters of the strategic decisions made. Get the strategy right, and agents will execute brilliantly at scale. Get it wrong, and they’ll execute the wrong thing at scale.

Architecture

Building the systems. Designing the workflows. Translating business goals into automated, multi-step agentic pipelines.

This is the job of system designers: AI ops, marketing operations, solutions architects, growth-oriented product managers. Agentic AI changes the constraint here. For the first time, complex, multi-step marketing workflows can be designed without writing code. People who understand marketing deeply, like the customer journey, the messaging architecture, and the campaign logic, can now build the systems that execute those things themselves. Workflows do not need to be routed through engineering every time. And this is where the leverage is. Architecture used to sit in the background: the plumbing marketing ops maintained so the “real” marketers could ship. Now every workflow an architect builds is execution capacity that didn’t exist before. The architect isn’t supporting the work. They’re multiplying it.

Architecture is also the exception to the “reshape, don’t reorg” rule. Most functions in an AI-native marketing org are about people you already have doing different work. Architecture is different. This is where you’ll need to add roles that weren’t on the org chart a year ago, like AI marketing ops, agent operations leads, and solutions architects who sit inside marketing instead of IT. The other pillars can absorb the shift by reshaping existing roles. Architecture requires specialists whose whole job is designing and maintaining the systems the rest of the org runs on. If you’re going to hire net-new for AI, hire here first. This is where the leverage of every other function gets built.

Curation

Encoding specialist knowledge into the system. Defining what “good” looks like in a way an agent can understand and reproduce, and creating what great looks like in the first place.

This is the taste-keeper job: product marketers, content directors, brand strategists, and copywriters. Curation is not prompt engineering. A prompt engineer writes instructions. A curator makes creative decisions. And an agent can’t make creative decisions. It can do some orchestration; it can definitely execute, but it can’t curate. That’s the line.

If Strategy answers what is being done and why, Curation answers how it should feel and what great looks like. That’s the creative standard, and it’s where craft lives in an AI-native org. There are two jobs inside this pillar, and both require human craft:

Creative: The breakthrough work

Curators come in many shapes: product marketers, content directors, brand strategists, senior copywriters, and campaign leads. What they share is the ability to make creative decisions an agent can’t.

Governance: Encoding the vision

Once the creative work exists, someone has to encode it into the system so agents can reproduce it at scale. Sticking with the product marketing example: if the creative is the positioning, the proof points, and the story, governance is the guardrails around how to use it, like which claims are approved, which competitors we name and which we don’t, how the messaging adapts by persona and stage, what language is off-limits.

Orchestration

Directing the agents that run campaigns. Setting goals, monitoring what’s working, adapting.

This is the system operator job that’s typically owned by campaign managers, integrated marketing, and demand generation. Orchestrators don’t produce the work. They direct the agents that do. In an AI-native org, everyone becomes a manager of humans, agents, or hybrid teams, and orchestration is where that shift shows up most clearly.

Execution

Producing the work. Drafting, formatting, scheduling, and first-pass research: high-volume, repeatable tasks that used to fill a specialist’s week.

Real-time intelligence

Sensing the market. Watching what’s happening across data, social sentiment, campaign performance, and the search results where buyers are now forming opinions, and surfacing it fast enough to act on.

The leverage: continuous sensing, not periodic reporting

If Execution is the leverage that comes from encoding your standards once and letting agents reproduce them at scale, real-time intelligence is the leverage that comes from sensing the market continuously instead of periodically.

Agents sense. Humans decide.

The critical word here is intelligence, not automation. This layer isn’t about agents making decisions for you. It’s about agents making sure the right human has the right signal at the right time.

How the pillars feed each other

None of these pillars work in isolation. The AI-native marketing org isn’t six functions running in parallel lanes. It’s a system where signal flows between them constantly, and the quality of each pillar depends on the others.

The context layer: The foundation of an AI-native marketing org

There’s one element of the AI-native marketing org that doesn’t map to a person or a role. It sits underneath all six pillars and determines how well they perform. This is the context layer.

The context layer has five components:

• Brand standards: the curated messaging, positioning, voice, approved terminology, and style guide that define how the brand shows up.
• The enterprise brain: the dynamic intelligence that usually lives in people’s heads, like how decisions actually get made, the reasoning behind past calls, and the institutional memory of what worked and what didn’t.
• Data: the customer, campaign, and market data that feeds the agents.
• The marketing tech stack: the connected systems agents draw on and write to.
• Market intelligence: current and accurate understanding of the market, competitors, and customers.

From pyramid to diamond: How the AI-native org chart reshapes

Today, most marketing orgs look like a pyramid: a small layer of leadership at the top, specialists in the middle, and a broad base of people doing repeatable, high-volume execution at the bottom.

AI absorbs that base. The entry-level, high-volume work that’s exactly what agents do well.

The pyramid doesn’t disappear. It reshapes into a diamond. Mass moves to the middle, where curation, architecture, and orchestration live.

Your first move isn’t hiring. It’s asking your existing team where they fit.

The temptation with a framework like this is to read it as a hiring plan: six functions, so go find six people. That’s the wrong first move. The people who can do this work are almost certainly already on your team.

When to hire

Once you know where your existing team fits, hire deliberately for the gaps. Architecture is usually the sharpest gap: as covered earlier, it’s the exception to the reshape-don’t-reorg rule.

The compounding advantage of getting this right first

The end state is agents coordinating with each other, planning, handing off, and escalating, while humans sit above the system setting goals, defining exceptions, and watching outcomes. The role is no longer doing the work. It is owning the result.

Here’s what I find most exciting: the teams that figure this out first are going to have a structural advantage that compounds over time. Every week a curator spends maintaining the system makes the system better. Every workflow an architect builds compounds. Every context layer a team builds becomes institutional knowledge that persists and scales across people, across campaigns, and across quarters.