Ian Kim
← POV
nov 16, 2026/4min

Toward a Coherent Marketing Operating Model

AI agents are scaling across marketing, yet the operating model caps the value they unlock. Today's model was built around what was hardest to get in the personalization era. Agents change that, and capturing their value calls for a coherent operating model.

Evolution of marketing operating model

Intro

Data, AI and now AI agents are being scaled across marketing to propel growth. In my experience with marketing transformation, especially in large enterprises, the biggest cap on that value is the operating model: who owns which decisions, how work moves between teams and what each team is measured on.

Harvard Business Review recently pointed out the same thing: marketing's constraint is the operating model, not the tools. Supermetrics found that 52% of marketing teams defer data strategy to an analytics team outside marketing.

So if agents take over execution, what should the operating model be designed around?

Every operating model is built around what is hardest to get

Teams form, and decisions get divided, around whatever is hardest to get at the time.

  • The agency era (1950s to 1990s). Creative talent and media access. Agencies and creative departments formed around them.
  • The martech era (late 1990s to early 2010s). Platform expertise and the ability to translate marketing ideas into engineering. Channel specialists and marketing operations formed around them.
  • The personalization era (mid-2010s to today). Unifying first-party data and activating it with precision. Data teams, CDP owners and personalization specialists formed around it.

The personalization era produced the model most enterprises run today: separate teams for email, paid media, lifecycle, data, creative and operations, each owning its channel's decisions, tools and KPIs. It optimized organizations for rapid technological change and became the standard across industries.

Its limit is the insight it leaves on the table. Questions that span channels, campaigns or brands have no natural owner. Which customers are four programs contacting this month? What did one campaign learn that the next should inherit? Where do two brands share the same customer? Each team holds part of the answer, and no team owns the question. Agents placed in this structure accelerate the station they sit in, and the model caps the rest of their value.

Agents commoditize execution and shift the value model

Agents now build segments, produce variants, launch campaigns, run analysis and report results. The specialist skill each channel team was built around is becoming abundant.

That is what "marketers will become directors" actually means. A marketer's value moves to setting intent, connecting signals and judging tradeoffs, which requires a new way of working, beyond a new title.

Platforms already reflect this. Coding agents work inside a harness that carries context across tools, and marketing is following. Hightouch co-founder Tejas Manohar recently described a marketing harness that lets AI see the work from one level higher and carry context across the team's tools. (Full disclosure: I work at Hightouch.) The operating model has to follow.

The leap to coherence

When agents execute and everyone moves fast, speed stops separating anyone. Two things likely will: how well a brand earns its customers' attention, now the scarcest resource on the customer's side, and how much the organization learns from every decision, the asset I described in collective decision intelligence.

I call the operating model that captures both coherence: marketing that decides with the whole customer in view and gets smarter with every decision. It is the model behind agentic scale, and it rests on seven design choices.

  • Decisions. Made at the customer level across channels and campaigns, including the choice to send nothing.
  • Learning. Carried beyond the single campaign, across channels and brands.
  • Exploration. Continuous, with marketers asking new questions of governed data and testing many ideas in parallel.
  • Measurement. Customer-level outcomes and learning, alongside channel KPIs.
  • Execution. Agents execute. Humans own intent, judgment and tradeoffs.
  • Context. Customer data, definitions and decision history shared by every marketer and agent, with guardrails set up front.
  • Talent. Marketers who form hypotheses, work with data directly, direct agents and connect signals across channels. Specialists keep their depth and encode it into shared context.

Early signs of change

Most marketing leaders I work with agree on the direction. The shift cuts across every team, so it rarely has a single owner, and the operating model moves far more slowly than the technology.

Pieces of a coherent model are already taking shape.

  • Integrated marketing. Brand, performance and lifecycle planned as one motion around the customer.
  • Unified sales and marketing. One revenue motion with shared data, goals and accountability.
  • Transformation towers. A dedicated team owns the shift across process, platform and talent.
  • Marketers directing agents. Marketers set intent and guardrails, and agents build, launch and optimize.
  • Insight-led exploration. Top teams use agents to find insights across campaigns that dashboards never covered.

These are signs of early adopters thinking ahead. Most are still pilots or partial moves, and few have become a full transformation of the operating model.

What this means for you

For two decades, marketing won by doing more, faster and with more precision in every channel. Agents now do that for everyone.

The next advantage is structural. If agents took over execution tomorrow, what would your teams be organized around?

Share