Governing the Agentic Enterprise: What Marketing Leaders Need to Know Now
The question used to be whether marketers should use AI. Now, that debate is over. The resulting tension is harder: On the one hand, the more unified and complete your customer data is, connected across touchpoints, channels and interactions, the more AI can deliver. But privacy and governance requirements raise the stakes: More data access requires tighter controls, clearer accountability and greater trust in how that data is used. Marketers today are grappling with competing priorities: Do we get value out of as much data as possible while still adhering to privacy needs? And how can we govern AI well enough to trust it at scale, and still move fast?
This week at Cannes Lions, we're launching the fifth edition of The Modern Marketing Data Stack report: Governing the Agentic Enterprise. I can see the arc of these reports and the marketing stack clearly. We started by asking whether organizations could unify their customer data. We moved on to asking whether AI could help them use it better. Now we’re diving into the tools and architectures that help you let AI agents act on your data while maintaining control and supporting compliance.
That is the defining challenge of the agentic era. As Scott Brinker, chief martec analyst, puts it in this year's report: "AI didn't reshape the stack mainly by replacing tools. It reshaped it by creating a new control plane above them."
The three forces are converging
Across past editions of this report, we have tracked three forces reshaping the modern marketing data stack: AI, privacy and data gravity. This year, the shift is more practical: Those forces are now showing up in the way work actually gets done.
AI is moving closer to execution and is beginning to shape workflows, trigger actions and influence decisions across the stack. That changes the role of the data foundation. When systems are making or recommending decisions, they need to work from governed data that teams trust.
Privacy is part of the same shift. As more decisions become automated, it’s even more important for consent and data usage rules to be built into the systems where data is accessed and action is taken.
That is why the governance framework becomes the through line. It gives teams a way to move faster without losing control. It makes clear when AI can act, when people need to stay involved and how outcomes are measured.
The convergence of AI, data gravity and privacy is changing how the modern marketing data stack is built, operated and governed.
1. The ROI of AI, and what governance actually enables
AI capabilities have expanded dramatically across the marketing stack. But the more consequential development of the past year goes beyond generative AI to the early emergence of agentic workflows: systems that do more than just assist with individual tasks and now coordinate sequences of actions, invoke tools, make decisions and optimize toward defined outcomes with human oversight and control.
The era of the agentic enterprise changes the operating model fundamentally. When AI can act on your behalf, the quality of your data foundation and, as this year's report emphasizes, the clarity of your governance controls determine whether those actions create value or introduce risk.
In the earliest days of AI experimentation, CMOs tried getting ahead by moving fastest or deploying the most tools. But the ones seeing the best results have paired specific use cases with quality, governed data, with the report highlighting examples from major organizations.
Fanatics illustrates what this looks like at scale. With billions of fan signals generated daily across sports, gaming and collectibles, and fragmented across separate business units, the company built FanGraph on Snowflake: a unified foundation stitching together a single view of more than 100 million fans. From there, they deployed Snowflake CoWork to put data directly in the hands of business teams. Marketing and operations stakeholders can now explore data, ask follow-up questions and surface insights without routing every request through a data analyst. With that governed foundation in place, they started building enterprise agents to automate insight-gathering and personalize fan experiences at scale. "We have a lot of data,” said Daniel Fox, Principal Product Manager at Fanatics. “We need to make sense of the data. Now we have the tools to actually do that and to empower and personalize the experiences of our fans."
As agents take on a larger role, one of the most important things to define first is accountability. Partners contributing to this year's report point to a consistent pattern: Agentic initiatives stall when ownership of decisions and measurement is unclear. The organizations making progress are the ones that establish governance early, with accountable owners.
2. Privacy: From compliance function to operating capability
If anything, the privacy landscape is more complex than it was a year ago. Enforcement has sharpened and consumer expectations have risen. As AI agents take on more autonomous roles, the surface area of potential risk grows.
The organizations I admire have stopped treating privacy as a compliance checkbox. They've embedded consent, identity and data usage controls directly into how their data is accessed and shared, especially in automated workflows. As AI agents act across more of the stack, you can't review every action in real time. Governance policies have to be built into the destination, not just the door.
And there's a deeper reason to get this right: consumer trust. In an era of increasing scrutiny, trust is one of the most valuable assets a brand can build and one of the hardest to rebuild once it's broken. First-party data, transparent consent and demonstrated value to customers remain durable advantages regardless of how the regulatory environment continues to shift.
3. Your data foundation is your agentic AI foundation
When we launched this report in 2022, the foundational argument was straightforward: Unify your marketing data in a single, centralized platform first, and deploy best-of-breed tools on top of that foundation. That principle is more important today than it has ever been. As AI-driven workflows span analytics, activation, measurement and collaboration, fragmented stacks built around duplicated data simply can't keep pace. As Scott Brinker says: "AI doesn't magically consolidate that ecosystem for you. It raises the stakes. The quality of your data foundation, semantics and operational controls becomes the main determinant of whether AI actually delivers value."
The defining question for marketing organizations in 2026 isn’t limited to what applications to include in their stack. It's whether the stack can support coordinated decisioning, governed data access and controlled automation across a complex environment. The stack has to reorganize around the data. And the data has to be governed.
What marketing leaders should do right now
First, there is no AI strategy without a data strategy, and no data strategy without governance. The organizations moving fastest with AI are the ones who built their foundation first — unified, consistent, with clear controls over how data is accessed, shared and acted upon. Governance isn't slowing innovation. It's what makes scalable, accountable AI-driven action possible.
Second, embrace composable thinking. The tools and categories mapped in this year's report number more than ever, but it doesn’t mean that organizations need to deploy most of them to succeed. The ones seeing results are assembling capabilities deliberately, choosing tools that work with their data where it lives, extending and replacing them without breaking the governance foundation underneath. In a world where the technology keeps changing beneath you, composability is how you stay agile.
As we launch our fifth report surveying the modern marketing data stack and its evolution, the thing I'm most certain of is this: the marketing leaders who will thrive in the agentic era are the ones who understand that governance and innovation are not opposites. The organizations closest to their data — and most deliberate about how it's governed — are the ones who will act fastest, and most confidently, when agents arrive in full force.
Inside the fifth edition of the Modern Marketing Data Stack 5th Edition
This year's report brings together perspectives from CMOs, marketing technologists and AI experts across our partner and customer ecosystem. It examines how marketing organizations are building the architectures, operating models and governance foundations required to make agentic AI a practical, business-critical capability. And of course, it maps the Leaders and Ones to Watch across the evaluated categories in the stack. These designations are grounded in Snowflake’s 12-month analysis of how customers use partner technologies. You’ll find a description of the full methodology in the appendix of the report.
Download the fifth edition of the Modern Marketing Data Stack report to explore the full findings.
And if you're here at Cannes this week, I'd love to continue the conversation in person.
