The conventional marketing department — sequential, siloed, perpetually behind schedule — is becoming obsolete. A new model built around autonomous AI agents and a shared intelligence layer is already delivering measurable returns: materials adapted 98 times faster, unit costs down 80 percent, click-through rates up 17-fold. Gulf brands are already in motion.
The conventional marketing department — sequential, siloed, perpetually behind schedule — is becoming a liability. As AI accelerates product development across the enterprise, the demands on marketing are expanding faster than the function’s inherited operating model can absorb.
The solution, argue three practitioners with front-line experience building AI-driven marketing systems, is a wholesale redesign: a new structure built for human-agent collaboration, anchored by a shared machine-readable intelligence layer they call the ‘brand code.’ Early adopters are already posting numbers that would have seemed implausible two years ago.
Writing in Harvard Business Review this month, the authors make the case for the ‘agentic marketing organisation’ — a model in which layered systems of specialised AI agents handle content creation, experimentation, distribution, and performance reporting at scale, while human marketers shift from execution to strategic direction and judgment.
The authors are Michelle Taite, former chief marketing officer of Intuit Mailchimp and previously VP of Global Marketing for Intuit QuickBooks, with earlier roles at Unilever and New Balance; John Winsor, executive fellow at Harvard Business School’s Digital, Data, and Design Institute, whose research focuses on AI, organisational design, and the future of work; and Will Fernandez, co-founder of Defyner, an AI-native marketing consultancy.
WHY THE OLD MODEL IS BROKEN
The authors’ diagnosis is structural rather than technological. AI has already compressed engineering and product timelines dramatically — research from Anthropic cited in the piece shows that software engineering now accounts for nearly half of all agentic AI activity, meaning product teams are shipping continuously rather than quarterly.
Marketing, however, still runs on the old cadence: cross-functional handoffs, status meetings, sequential approvals. Faster AI tools inserted into that architecture do not resolve the underlying friction; they only make the bottleneck more visible.
“The issue isn’t the tools. It’s the operating model,” the authors write. “Most marketing workflows remain constrained by sequential processes and siloed systems, making it impossible to operate at the speed the business now demands.” No CMO, they note with characteristic directness, sets out to become a bottleneck. Yet that is what is happening.
Companies such as HubSpot and AWS have begun implementing the agentic model. Across these deployments, the authors report marketing materials adapted up to 98 times faster, unit costs reduced by 80 percent, and click-through rates increased up to 17 times.
Research from BCG has demonstrated these benefits at scale, finding that organisations embedding agentic AI into marketing workflows can achieve up to a threefold increase in ROI, campaign speed, and content volume.
“The issue isn’t the tools. It’s the operating model. Most marketing workflows remain constrained by sequential processes and siloed systems.”
THE BRAND CODE: A SHARED INTELLIGENCE LAYER
At the core of the proposed model is what the authors call the ‘brand code’ — a machine-readable knowledge base encoding brand strategy, customer insights, product information, and business rules in structured formats such as taxonomies, prompt templates, decision trees, and tagged datasets that both people and AI agents can directly reference and act upon.
Think of it, the authors suggest, as the always-on onboarding documentation the entire organisation — human and machine — needs to function effectively. The brand code is not static: as campaigns run, performance data feeds back into the system, continuously refining messaging, audience definitions, and decision logic. It also solves a problem endemic to large organisations — the institutional knowledge lost when employees leave. By encoding that knowledge systematically, the brand code makes it persistent and shared.
A PLATFORM OF LAYERED AGENTS
Above the brand code, the authors describe three functional layers. An execution layer houses specialised agents, each responsible for a discrete task: content generation, localisation, experimentation, or deployment.
An orchestration layer coordinates these agents dynamically, managing dependencies and routing outputs as project plans, status meetings, and manual handoffs once did. An interface layer gives human marketers a single surface — embedded in familiar tools such as Slack, WhatsApp, or Teams — through which to set intent, review outputs, and make decisions when the system escalates them.
Together, the layers transform a fragmented marketing stack into a coordinated system governed by shared intelligence rather than individual judgment calls. What was once handled through project plans and status meetings is instead managed dynamically by the system.
FIVE WORKSTREAMS, REDEFINED
The authors map marketing activity onto five agentic workstreams. In intelligence and ideation, agents continuously synthesise market signals and performance data, replacing the front end of the planning cycle with structured, prioritised briefs ready for human evaluation. In content creation, agents generate and adapt material at scale within defined parameters, freeing marketers from production cycles. In research and testing, agents design and execute experiments continuously rather than episodically. In distribution, agents handle adaptation, scheduling, and deployment across an increasingly fragmented channel landscape. In performance and reporting, agents monitor results in near real time, feeding learnings back into the system rather than waiting for retrospective reviews.
Across all five workstreams, the human role shifts from execution to direction — setting strategic intent, evaluating system outputs, and determining when judgment is required that the system cannot provide.
THE MARKETER’S IDENTITY CRISIS
Perhaps the most candid section of the HBR piece addresses the human dimension of the transition. Many marketers have built a professional identity around the act of production — writing copy, shaping assets, managing handoffs. In an agentic model, that satisfaction migrates upstream, from doing the work to directing the system that does it. The authors acknowledge this is not purely additive: it requires letting go of deeply ingrained instincts.
Organisations must consequently change how they hire — seeking people who think in workflows and systems rather than tasks — and how they manage, shifting from reviewing deliverables to reviewing whether feedback loops are functioning and whether the brand code remains current and accurate. The marketers who adapt fastest, the authors argue, will not necessarily be the most technically proficient; they will be those who can recognise quality in context and shape how the system evolves based on their observations.
THE GCC IS ALREADY IN MOTION
The agentic shift the HBR authors describe is not an abstract future scenario for Gulf marketing leaders. It is already unfolding across the region — and in some respects, the GCC is moving ahead of global benchmarks.
Amer El Hajj, CEO of WPP Media MENA, recently told Communicate: “The Middle East is not just a fast-follower anymore. With the rollout of in-store AI sensors and bilingual agentic commerce, the GCC is setting a global benchmark for how digital and physical retail media should work together,” Amer El Hajj, the CEO of WPP Media – Mena, said in 2026.
Gulf enterprises are also investing at the scale the HBR framework requires. Communicate’s coverage of the AI enterprise blueprint in the Gulf found that e& (Etisalat) reported a 20 percent increase in enterprise revenues, driven by cloud and AI-managed services for both government and private sectors. The Chalhoub Group is deploying a group-wide AI strategy to unify disparate data points into a cohesive, highly personalised customer experience — structurally analogous to the brand code the HBR authors describe.
Sophie Simpson, Managing Director of Ruder Finn Atteline MENA, told Communicate: “LLMs and AI are a core part of how we market products and build reputation. Our dedicated RF Tech Lab helps brands and people understand how they show up inside AI answers and structure content so it’s findable, useful and citable.”
THE CULTURAL NUANCE PROBLEM: WHERE THE GCC PUSHES BACK
Not every dimension of the HBR model transfers without friction to the MENA context, and Gulf practitioners are candid about the limits.
Mohamed Itani, CEO of United Foods Company, told Communicate that while AI agents can optimise bids, timing, and targeting, achieving even 80 percent automation feels optimistic in this region. Brand tone, cultural nuance, and media mix still require a human lens, he argued: machines can scale decisions, but brand stewardship still sits with people who understand context.
Imen Chatti, Founder, INTERCULT BRANDS, an international branding and design consultancy, Paris, France, said:
“The GCC market is highly fragmented. The issue isn’t access to technology; it’s stitching identity and measurement together. The GCC purchase journey is genuinely non-linear. The problem is that activity is spread across walled gardens, modern trade, traditional trade, retail platforms and messaging apps, creating disconnected data flows.”
She added that concrete signals show the scale of the issue:
“In the UAE, 63 percent of marketers say they are concerned about brand suitability on social platforms. Operationally, around one-third say managing campaigns across multiple channels and calculating ROI are among their biggest challenges.”
