There's a cloud hanging over the AI era, and I'm not hearing nearly enough people talk about it. Every marketer I speak with is chasing the same vision: AI-driven agents that can build, execute, and optimize campaigns across every channel in real time, no bottlenecks, no handoffs, no delays. Honestly, I get it. That vision is real, the technology exists, and the budgets are moving. But I keep coming back to… why does it still feel so far out of reach? Because over the last decade, marketing stacks have grown incredibly powerful and incredibly fragmented. DSPs, CDPs, CRMs, clean rooms, identity layers, measurement tools, each one doing its job, but mostly designed to operate on its own terms, not as part of a connected whole. On paper, everything talks to everything. In practice, it rarely does.

Here's the part that doesn't get said enough, AI doesn't just need data … it needs to act on it. An agent running a live campaign has to read from one platform, write to another, trigger an update in a third, and report into a fourth, in real time, without waiting on a nightly batch job or a manual export. That's a completely different demand than anything we've put on these systems before. And most stacks today weren't built for it. The problem is the gap between what AI can theoretically do and what the ecosystem can support.

What we're really talking about is interoperability. The industry has made genuine progress on data standards, identity, measurement, privacy compliance, and I don't want to minimize that. But interoperability, the ability for platforms to connect, communicate, and act together in real time, has lagged badly behind everything else. Some of that is competitive dynamics. Platforms have little incentive to make it seamless for their users to work with rivals, so they don't. Some of it is the API landscape, which is fragmented, inconsistent, and constantly shifting underneath the teams trying to build on top of it. Some of it was that the urgency wasn't there yet, so no one forced the issue. AI made it urgent overnight. I've watched organizations invest heavily in models and agents, get genuinely excited about what's possible, and then run straight into the wall of a stack that simply cannot move fast enough to support them. The intelligence is there, but the infrastructure isn't. And until we're honest about that gap as an industry, we're going to keep selling a vision we can't fully deliver on.

An agent that can plan but can't execute isn't an agent, it's an expensive research tool. The value is in the action layer, not the model layer. And that action layer only works as well as the connections beneath it.

So what actually has to change? Vendors need to treat interoperability as a core product commitment, not a checkbox on a sales sheet. And buyers need to evaluate their stacks on the performance of the whole system, not just what each tool does in isolation. The agentic marketing era won't be decided by who has the best model, it'll be won by the teams and ecosystems that can act on data quickly, reliably, and at scale. We’ve got the intelligence to execute, the real work is enabling it to act.