Why the World’s Largest Brands Are Building Systems — and What You Can’t Afford to Get Wrong
The Question Has Changed
Somewhere in the last 18 months, the nature of the question changed.
For most of the digital era, marketing organizations asked some version of the same thing: how do we use data, technology, and automation to market more effectively? It was a reasonable question. Entire industries emerged to answer it.
But as AI has moved from experiment to operating reality, the leaders we speak with are asking something different. They are less interested in the next tool or the next platform. They are trying to understand something harder: whether they own the capability to generate intelligence for themselves, or whether that capability has been quietly eroding for years.
At first glance, that sounds like a technology concern. It isn’t. Every organization has infrastructure it would never outsource entirely — finance owns its books, supply chain owns its inventory — because those assets determine how the business operates. Marketing has historically been different, delegating not only execution but the systems that interpret performance, define audiences, structure data, and shape decision-making.
As AI becomes embedded into planning, content, measurement, and commerce, those systems are becoming part of the intelligence layer of the enterprise itself. The question is no longer whether they are efficient. The question is whether the organization has the capability to own, govern, and compound what they produce — or whether that capability was quietly delegated away.
The issue is not that AI is moving fast. The issue is that every institution around the brand is reorganizing around intelligence at the same time.
What It Costs to Not Build
The consequences of not building owned intelligence rarely appear dramatic at first. The problem emerges as friction. Decisions take longer than they should. Different teams arrive at different answers to the same question. Reporting becomes more sophisticated, yet confidence in the numbers declines. Agencies become difficult to replace. New technology gets added while old technology never seems to disappear.
Individually these problems appear manageable. Collectively they point to the same source: a capability gap, not a technology problem.
Over the last decade, organizations invested heavily in platforms designed to improve targeting, measurement, personalization, and analytics. Each investment was rational. The result is that many now operate dozens of systems that were never designed to function as a coherent whole. Taxonomies drift. Definitions vary. Teams spend increasing amounts of time reconciling information rather than acting on it. The organization accumulates complexity faster than it accumulates intelligence. AI amplifies this dynamic in both directions — organizations with governed data and consistent workflows see meaningful acceleration; organizations with fragmented systems discover that automation scales inconsistency.
Measurement creates a specific version of the same problem. Every executive has experienced the moment when multiple reports describe the same performance differently. Marketing reports one answer. Finance reports another. The agency reports a third. What appears to be a reporting issue is a governance issue. When the underlying logic is inconsistent, confidence erodes — not just in the numbers, but in the organization’s ability to know what is true.
There is a version of this that has nothing to do with agencies or platforms at all. Most organizations have never prioritized understanding how their own work gets done — not in the sense of a process diagram that lives in a handbook nobody reads, but operationally. How does a brief actually move from strategy to execution? Where do approvals stall? Which steps exist because of workarounds nobody has revisited? That institutional knowledge lives in people’s heads, never written down because it never needed to be. When AI needs to operate in that environment, it finds nothing to work with. The dependency here is not on an external partner. It is on individual people carrying knowledge the organization has never chosen to capture — and that is its own form of fragility.
The most expensive consequence rarely appears on a balance sheet. A brand building its intelligence capability now will have years of compounding learning that a brand starting later cannot quickly replicate. The gap is not visible in a single quarter. It becomes apparent when one organization is operating from accumulated intelligence and another is still establishing the foundations.
The Signal in the Capital
One of the easiest ways to understand where an industry is headed is to stop listening to what it says and start watching what it buys.
The world’s largest brands are spending billions building systems designed to centralize intelligence. P&G has built Consumer 360 — a unified view of consumers and shoppers that connects data, insights, and decision-making across the enterprise. Nike has brought critical measurement and performance functions closer to the business. L’Oréal has spent a decade transforming itself into what it now calls a Beauty Tech company, with proprietary systems connecting consumer data, content creation, AI-enabled workflows, and product innovation. Coca-Cola, Unilever, and Heineken are pursuing different strategies but arriving at the same conclusion. The scale of those investments reflects their competitive context and organizational maturity — not a template every brand should replicate, but a direction of travel every brand should understand.
The pattern is not limited to brands. Publicis acquired Epsilon, then Lotame, then LiveRamp. WPP acquired InfoSum and launched Open Intelligence. Omnicom merged with IPG, assembling one of the largest collections of consumer intelligence assets in the industry. These are not creative acquisitions or talent acquisitions. They are intelligence acquisitions — made because the most sophisticated organizations in marketing believe that ownership of identity, data collaboration, measurement infrastructure, and AI-enabled decision systems will become increasingly valuable.
What P&G, Nike, L’Oréal, Publicis, WPP, and Omnicom have all recognized is that intelligence is beginning to behave like a strategic asset.
Historically, competitive advantage came from assets that were difficult to replicate: manufacturing scale, distribution networks, retail relationships, media buying power. Today’s emerging advantage looks different.
Technology depreciates. Campaigns end. Media spend disappears the moment it is spent.
Intelligence compounds.
Every interaction creates a signal. Every signal improves understanding. Better understanding produces better decisions. Better decisions create better experiences. Better experiences generate more signals. The cycle reinforces itself.
But only if the organization owns the system that captures it.
What Ownership Actually Looks Like
Ownership is not an insourcing strategy. The brands most often cited as leading this transformation all operate within complex ecosystems of agencies, platforms, and technology partners. The question of what to build internally versus what to partner on is a strategic decision every organization has to make for itself, shaped by capability, culture, and competitive context.
What that decision has to include is an honest assessment of capability — what the organization has built, what has atrophied, and what has never been prioritized in the first place.
Most brands arrived at their current state not through bad decisions but through accumulated convenience. The agency built the dashboard because they had the data. The taxonomy lived in their system because they set it up first. The attribution model was their proprietary methodology because they proposed it and it worked well enough. Each individual choice was defensible. The aggregate is a capability gap — the ability to define audiences, evaluate performance independently, and understand how the organization’s own marketing actually works has quietly migrated outside the brand’s direct control. The problem was not that agencies did something wrong. In many cases they did the work well. The problem is that brands allowed critical capabilities to atrophy because someone else was handling them.
That gap wasn’t visible while the model was working. It becomes visible when the model changes — and AI is changing the model. Agencies are now offering to be the AI operating system: the intelligence layer, the agentic workflows, the data infrastructure, the measurement logic, all integrated. It is a compelling offer that solves real problems. But it carries the same structural risk as the first wave, one level deeper. The first wave cost brands access to intelligence. The second wave risks costing them the capacity to generate it — because rented capability doesn’t compound.
There is a version where both things are true simultaneously. When the brand owns the foundation — the data, the taxonomy, the measurement logic, the workflow definitions — and the agency’s AI capabilities connect to that foundation, the relationship compounds in both directions. Brand infrastructure and agency capability connecting as peers, each making the other more valuable. That is the architecture worth pursuing, and the role of an independent advisor is not to advocate for a particular outcome but to ensure the foundation is solid enough that any partnership delivers durable value.
But owned capability is not only about what faces outward — consumers, campaigns, performance. It is equally about what faces inward.
Most organizations have never prioritized understanding how their own work gets done. Not in the sense of an org chart or a process diagram — but operationally. How does a brief move from strategy to execution? Where do approvals stall? Which steps exist because the data wasn’t accessible, the systems didn’t connect, or someone made a decision years ago that nobody has revisited? That institutional knowledge lives in people’s heads, accumulated informally over years of practice. It is never written down because it has never needed to be.
AI needs it written down.
Before an agent can help, the work has to be legible. The organization has to understand its own processes clearly enough to describe them, govern them, and eventually improve them. That is a form of intelligence infrastructure entirely within the brand’s control — and entirely dependent on whether the brand has ever chosen to prioritize it.
This is what makes the ownership question broader than it first appears. Not only who holds the data or controls the measurement methodology — but whether the organization understands itself clearly enough to compound that understanding over time. The brands building toward an AI-enabled future are doing both simultaneously: establishing owned capability in the external intelligence domains and developing the operational clarity that makes internal AI deployment actually work.
Neither happens by accident. Both require a choice to prioritize them. And both require the same foundational discipline: establish governance before scaling automation. The organizations that will create lasting value from AI are the ones that have determined who governs data, they’ve aligned around common taxonomies, defined how performance will be measured, and documented how work flows — before they scale the tools. None of that is exciting work. It is almost always the difference between intelligence that compounds and complexity that accumulates.
Intelligence infrastructure is not a transformation you complete. It is a maturity you pursue — and the most valuable first step is an honest assessment of what you have, what has atrophied, and what was never built in the first place.
The Direction of Travel
The conversation has changed names repeatedly — data, customer experience, digital transformation, artificial intelligence. The underlying challenge has not. Organizations are trying to determine how they will learn, adapt, and make decisions in an environment that is becoming more complex, more automated, and increasingly shaped by intelligent systems.
The challenge is not access to AI. Access is rapidly becoming universal. The challenge is whether the systems surrounding that AI — and the processes that govern how the organization actually works — are capable of learning and improving over time. That will be determined less by any individual technology and more by the quality of the foundations beneath it.
Intelligence behaves differently from the assets that preceded it. Technology can be purchased. Platforms can be replaced. Agencies can be changed. Intelligence, once accumulated and governed effectively, becomes increasingly difficult to replicate — because it is embedded within the operating system of the organization itself.
The organizations making the most progress started with a single honest conversation — not about technology, not about agencies, not about budget. About themselves. What do we actually know? What have we stopped knowing? What have we never known? That conversation is harder than it sounds. It requires setting aside the comfort of activity and asking whether the activity is building something that lasts.
The future this document describes is being built right now, by organizations that made one deliberate decision: to treat their intelligence infrastructure as something they own, not something they rent. That decision, more than any technology, is what separates them.
That choice is available to every organization reading this. The honest question is whether you are ready to make it.
This is not a transformation you complete. It is a maturity you pursue. And the most important step — at any stage — is knowing honestly where you are.
