The AI-native marketing function isn’t a team using AI. It’s a team designed around it.

You and I both know, almost every marketing team is using AI these days: They’re writing prompts, summarising research, creating first drafts, generating images, analysing campaigns, maybe even building copilots or experimenting with agents.

But that doesn’t make them AI-native. It makes them a traditional marketing organisation with some very cool and powerful new tools. I’ve caught myself doing it: reaching for AI to speed up a process before asking the more uncomfortable question, should this process exist at all? And I think we need to start recognising the difference.

The real opportunity with AI isn’t about making the marketing function today 20% faster. It’s to ask whether we would design that function in the same way at all if we were starting from scratch with what we know now.

We are automating yesterday’s organisation

This is the trap I see increasingly. We take an existing process, built around existing roles, existing team boundaries, existing approval chains and existing assumptions about how long work should take, and add AI somewhere in the middle. The content team uses AI to draft faster. Creative uses it to produce concepts. PR uses it to research journalists. Demand generation uses it to personalise campaigns. Leaders use it to summarise meetings.

Useful? Absolutely. Transformational? Not necessarily. And the companies who are caught in this trap are seeing their bottom line increase with exponential token costs, without any real productivity gains on the front end.

The data bears this out.

McKinsey surveyed more than 500 marketers globally in 2026. Nearly 60% said they were already using AI multiple times a week. But fewer than 10% had started capturing value across end-to-end marketing workflows.

Even more telling: only 28% described their organisations as fundamentally rewiring their marketing teams and workflows around AI. Most were still layering AI on top of the way work had always been done.

That, for me, is the gap that matters: lots of AI activity, not nearly enough actual transformation.

And I suspect it explains why so many organisations simultaneously believe AI is revolutionary and are slightly disappointed by the results. After all, a shiny agent bolted onto a terrible workflow is still a terrible workflow. We’re giving people Formula 1 engines and asking them to drive the same pot-holed roads.

AI-native starts with the work, not the tool

Over the past year, I’ve spent a lot of time thinking about what an AI-first marketing organisation should actually look like. And just to be clear, it absolutely does not mean the same old org charts now with robots and AI licenses. This isn’t about the tools we should buy (after all, a license is not an operating model), or whether everyone has access to your chosen LLM. It doesn’t mean that all your marketers have completed their AI training (although AI skills are fundamental and I’ll come back to that other day). Instead, the more interesting questions are more like:

• What work should be uniquely human?
• What work should AI augment?
• What work shouldn’t require a human at all?
• And then: With all of the above in mind, what should the team look like?

That leads you somewhere very different. Instead of looking at an org chart and asking, “where can AI help these people?”, you start with the outcome and ask;

What combination of human judgement, specialist expertise, data, automation, and agents is the best way to achieve it?

It may sound like semantics. It isn’t. Suddenly, you’re not designing jobs. You’re designing how the work gets done. And I think that is where AI-native marketing really begins.

Microsoft describes a similar shift in its Work Trend Index: away from rigid organisational charts and towards more fluid, outcome-driven “work charts,” where humans and agents are assembled around the work that needs to happen. Its research found 82% of leaders believe this is a pivotal year to rethink core elements of strategy and operations, while 46% already say their organisations are using agents to automate entire workstreams or business processes. And FYI, Marketing is one of the top areas receiving investment.

That starts to look less like “using AI,” it starts to look more like redesigning the company.

The role of the marketer changes too

There’s another implication here that I don’t think we talk about enough. If AI-native marketing is designed around humans and intelligent systems, then being excellent at marketing will increasingly mean being excellent at orchestration. In other words, I’m increasingly less interested in who can produce the most, and more interested in who can make the smartest calls:

• who can frame the problem best
• who knows what great looks like
• who can brief a human or an agent extraordinarily well
• who understands the brand deeply enough to reject something that is technically competent but strategically wrong
• who can connect customer insight, commercial context, creative judgement and data
• who knows when to automate, and critically, when not to.

This is why I don’t subscribe to the idea that AI makes human expertise less important. I think it makes real expertise much more visible.

AI can generate ten headlines in seconds. Knowing which one your brand should say requires human judgement. AI can produce a competitor analysis overnight. Knowing which competitive threat should change your strategy requires human judgement. AI can generate thousands of pieces of content. Knowing what is actually worth saying requires human judgement.

When everyone can make more stuff, faster, the real differentiator becomes knowing what’s actually any good. Taste, judgement, curiosity and strategic clarity become more valuable, not less.

And for me, that is one of the most interesting contradictions of AI-native marketing: The more capable the machines become, the more important the genuinely human parts of marketing become too.

Small teams may suddenly have very big capabilities

This also challenges one of the oldest assumptions in marketing leadership: that capability scales roughly with headcount. AI starts to break that relationship.

Microsoft’s research talks about closing the gap between business demand and human capacity through “digital labour.” Eighty percent of workers in its global study said they lacked the time or energy to do their jobs (something I think we can all relate to), while 82% of leaders expected digital labour to expand workforce capacity within 12–18 months.

I find this idea particularly interesting for marketing. Historically, wanting to do more usually meant needing more people: More campaigns? More campaign managers. More content? More writers. More markets? More localisation. More insight? More analysts. More creative? More designers.

AI changes the economics.

And this is where it gets interesting. A ten-person marketing team can suddenly start behaving like a much bigger one, without actually becoming a much bigger one. But (and this matters) that doesn’t mean you can replace ten people with one person plus Claude. That’s just cost cutting dressed up as transformation. The more ambitious question is:

What could ten brilliant marketers achieve if each of them suddenly had access to capabilities that previously required another fifty people?

That is where this becomes a growth conversation rather than an efficiency conversation. And marketing leaders need to defend that distinction. McKinsey found marketers themselves see AI relatively evenly as a driver of growth and efficiency, but believe the C-suite tends to see it much more through the productivity and cost lens. If we allow AI transformation to become solely a headcount equation, I think we will squander a large part of its potential.

AI-native does not mean human-lite

There is also a danger that AI-native becomes shorthand for automating everything because we can. I really hope it doesn’t. There are parts of marketing I actively don’t want to optimise away:

• The debate with a brilliant creative over the next campaign.
• The conversation with a customer where you suddenly hear the problem differently.
• The gut instinct that tells you a perfectly researched message somehow just doesn’t feel right.
• The strange idea that arrives because two people with very different experiences happen to be in the same room.

That unique human messiness is often where the interesting work happens. And marketing and brand teams need to be careful here to protect it.

Deloitte’s research found brand authenticity had become a growing concern among marketing leaders using generative AI, cited by 52% of respondents, alongside concerns about creativity, privacy, and reduced human interaction.

AI is making average marketing ridiculously easy to produce, so we’re about to get a lot more of it. And this will make being genuinely distinctive a much harder, and much more important job. AI-native cannot mean surrendering the brand to the machine.

It means using machines to create more space for the humans to do the things machines cannot easily replicate: imagination, empathy, judgement, provocation, storytelling, and taste.

So what does an AI-native marketing function actually look like?

I don’t think anyone has the definitive org chart yet, and that’s partly the point. We are still trying to discover it, but I do increasingly believe it has a few characteristics:

• It organises work around outcomes rather than departmental handoffs.
• It treats agents as part of the workforce architecture rather than an interesting side project.
• It gives marketers permission to redesign processes instead of merely accelerating them.
• It makes AI fluency part of everyone’s role rather than creating a tiny priesthood of “AI people.”
• It values people who can connect disciplines rather than defend silos.
• It has very strong human ownership of brand, judgement, ethics, and quality.

And it relentlessly asks one question:

If we were designing this today, knowing what AI can now do, would we still design it in the same way?

Sometimes the answer will be yes, but increasingly, I suspect it won’t.

McKinsey estimates generative AI alone could increase the productivity of marketing spend by 5–15%, representing hundreds of billions of dollars of potential value globally. But its more recent research makes an important point: the bigger prize comes from rewiring workflows around human-AI collaboration, not simply introducing more AI tools. That feels right to me. Because the organisations that win this transition will not be those with the longest list of AI licences. They will be the ones brave enough to reconsider how the work gets done.

We spent the first phase of generative AI asking: “How can AI help my marketing team?” The question for the next phase is much more interesting: “If AI was part of my marketing team from day one, how would I do things differently?”

That is the conversation I think marketing leaders should be having now, because this isn’t really a conversation about AI tools, it’s a conversation around what a marketing team does, and whether the one we have is still the one we need.

And an AI-native marketing function isn’t just a team using AI, it’s a team designed around what humans and AI can do together.

Article references

• McKinsey — From anxiety to advantage: A marketing organization that thrives with AI
• Microsoft — 2025 Annual Work Trend Index: The Frontier Firm is born
• Deloitte — CMO Insights: Trend 2 — The Technical Renaissance

Written by Emma Hunter, assisted by AI.