The 1.1% Problem: Why Most AI Governance Is Just Theater
1.1%
That’s the share of marketing organizations that have figured out both how to measure their AI investments and what return they expect from them. Not one in ten. Not one in fifty. One in a hundred. The Association of National Advertisers published these numbers in January 2026, surveying the state of AI adoption across the marketing industry (MarTech’s coverage here), and that single data point tells you more about where we actually are than any vendor keynote or analyst prediction from the past two years.
The rest of the survey looks impressive on the surface. About 76.6% of organizations now have AI policies in place, up from 55.3% just a year earlier. Nearly 89% plan to increase their AI spending. Two-thirds say they’d maintain that investment even if the economy cratered. Reading the toplines, you’d think the industry has its act together.
It doesn’t.
The Guardrails Came Before the Road
Here’s what happened, and it follows a pattern that anyone who’s worked in enterprise technology will recognize immediately. AI burst into the mainstream. Executives saw headlines about compliance risks, data privacy lawsuits, and deepfake scandals. The natural response was to establish control. Form a committee. Write a policy. Announce governance.
And that’s exactly what they did. Over half of marketing organizations now have cross-functional AI steering committees. Three-quarters have formal policies. The governance infrastructure looks robust.
But governance is supposed to follow strategy, not replace it. When you implement a CRM or a marketing automation platform, you start with the business case. What outcomes do we need? What processes change? How do we measure success? Then you build the guardrails around the plan. With AI, the industry reversed the sequence entirely. They built guardrails around a road that doesn’t exist yet.
The ANA data makes this painfully clear. Nearly half of organizations have no formal AI planning horizons at all. Over 71% haven’t established ROI targets for their AI investments. They can tell you what their people aren’t allowed to do with AI. They can’t tell you what they’re trying to accomplish with it.
This is governance theater. The appearance of control masking the absence of strategy.
We’ve Seen This Movie Before
If you’ve spent any time in the martech trenches, this pattern should feel uncomfortably familiar. For the better part of a decade, marketing organizations accumulated tools at a staggering pace. The average martech stack ballooned from a handful of platforms to hundreds. Vendors promised efficiency, personalization, and competitive advantage. Organizations bought in.
The result? Tool utilization rates sitting at roughly 42%. Disappointment hovering around 55%. Billions spent on capabilities that most teams barely touch, not because the tools don’t work, but because nobody planned for how they’d fit into actual workflows or drive actual outcomes.
AI is following the exact same trajectory. It’s just moving faster.
When the ANA asked marketers what primary value AI delivers right now, the top answer was time efficiency at about 61%. The near-term priorities cluster around tactical execution: content creation, workflow efficiency, personalization. These are operational improvements. They make existing processes faster. They don’t make organizations more strategic, and they don’t require governance committees to manage.
The uncomfortable truth is that most organizations are using AI to do the same things slightly quicker, governing that usage with policies designed for a level of sophistication they haven’t reached, and spending aggressively on capabilities they can’t measure. This isn’t adoption. It’s a dressed-up version of the same accumulation trap that martech fell into, except this time the tools are more powerful and the spending is accelerating faster.
The Leadership Hallucination
The ANA research segments the marketing workforce into behavioral groups based on experience. The largest group, about 61% of the workforce, are what they call “strategic governors,” marketers with 12+ years of experience who should, theoretically, be guiding AI adoption with the hard-won wisdom of previous technology cycles.
These strategic governors report the highest confidence in their organization’s AI journey. They also report being the most overwhelmed by the pace of change. Sit with that for a moment. The people most confident in the direction are simultaneously drowning in the speed. Confidence without a strategic plan isn’t leadership. It’s hope wearing a suit.
And when you compare leadership views to practitioner views, the gap widens further. Senior leadership shows about 52% optimism. Practitioners report nearly 30% anxiety. Executives see AI as a strategic opportunity. The people actually using it every day experience it as an operational burden. Without shared planning frameworks that connect the vision at the top to the execution at the bottom, these groups are working toward fundamentally different goals and neither side fully realizes it.
The strategic governors are supposed to be the bridge. Instead, they’re standing in the middle of the gap, confident they understand both sides, overwhelmed by the impossibility of connecting them, and governed by policies that address neither problem.
Agents Are Coming. Strategy Isn’t.
Here’s where this shifts from a measurement problem to an operational crisis. According to the ANA data, 37.4% of organizations plan to deploy agentic AI within the next six months.
If you’ve been following the agent conversation, you know the promise: autonomous systems that can execute multi-step tasks, make decisions within defined parameters, and operate with minimal human oversight. The governance implications are enormous. Agents don’t just assist human workflows. They create their own.
But agents are amplifiers. They amplify strategy when it exists. They amplify chaos when it doesn’t. An agent operating within a well-defined strategic framework, with clear objectives, measurement protocols, and oversight structures, can be transformative. An agent deployed into an organization that can’t articulate what it’s trying to achieve with AI, that has no ROI targets, no planning horizons, and no measurement sophistication? That agent will execute bad processes with ruthless efficiency. It will automate confusion at scale.
And the governance policies these organizations have written? They were designed for a world where humans use AI tools. They weren’t built for a world where AI systems take autonomous action. The steering committees, the usage policies, the compliance frameworks: none of that infrastructure was designed for agents, and the organizations deploying them in six months haven’t updated it.
This is what encoding strategic failure looks like. You take an organization that skipped from governance to deployment without stopping at strategy, hand it autonomous systems, and watch it bake its lack of direction into automated workflows that run 24/7.
What Strategy Before Governance Actually Requires
The prescription isn’t complicated. It’s just uncomfortable, because it requires admitting that the governance work most organizations have done isn’t wrong, it’s just premature.
Strategy before governance means starting with the question most organizations have skipped: what does your AI investment need to make true about your business in 18 months? Not “what tools should we buy” or “what policies should we write.” What has to change about your operations, your customer experience, your competitive position? If you can’t answer that clearly, then every policy you’ve written is decoration on an empty building.
From there, the sequence matters. Planning horizons come before tool selection. ROI targets come before budget allocation. Measurement frameworks come before scaling investment. And governance wraps around all of it, not as a starting point, but as the structure that keeps execution aligned with intent.
The cross-functional steering committees that over half of organizations have formed aren’t useless. They’re just pointed in the wrong direction. A committee that governs without planning becomes a review board. It slows adoption without improving outcomes. Redirect those committees toward strategic planning, defining use cases, establishing success criteria, building measurement capability, and you’ve turned overhead into infrastructure.
The 1.1% who’ve achieved both measurement sophistication and high ROI expectations didn’t get there by writing better policies. They got there by doing the foundational work that policies alone can’t replace. They defined what success looks like before they defined what failure isn’t allowed to look like.
The Window Is Closing
There’s a temptation to treat this as an industry maturity issue, something that will work itself out as organizations gain more experience with AI. I don’t think that’s realistic. The martech parallel is instructive here: the industry had over a decade to mature its approach to marketing technology, and utilization rates never recovered. Organizations that accumulated tools without strategic frameworks didn’t eventually figure it out. They just accumulated more tools.
AI is moving faster, and agents compress the timeline further. Organizations that deploy autonomous systems without strategic foundations aren’t going to have the luxury of gradual correction. Once you’ve automated processes built on no clear strategic logic, unwinding that automation is harder than building it right in the first place.
The 1.1% figure should be a wake-up call, but it will probably be a footnote. Most organizations will read the ANA data, note the impressive adoption numbers, and continue building governance around a strategy that doesn’t exist. They’ll form another committee, write another policy, and declare another victory.
The question for every organization deploying AI right now isn’t whether your governance framework is robust enough. It’s whether you have a strategy worth governing.