Here is the honest truth about where most AI conversations go wrong in marketing: they start with the technology instead of the work.
I came across a conversation between data strategist Juan Sequeda and a podcast host, recorded mid-flight over the English countryside, that put language to something I have been seeing in my own client work. The framing was so clean I want to break it down for every CMO thinking through their AI strategy right now.
The three levels of work
The host introduced a framework that stopped me in my tracks. When you talk about work, you are operating at one of three levels:
- Inputs: the tools and technologies you use. “We use AI for that.” “We run Spark.” “We migrated to Databricks.”
- Outputs: the things you build. “We launched a chatbot.” “We built a dashboard.” “We produced a content calendar.”
- Outcomes: what actually changes for the business. “We reduced churn by this amount.” “We increased qualified pipeline.” “We cut the cost per acquisition.”
The best marketing teams have always operated at the outcome level. They understand business context, they tie their work to revenue and retention, and they earn a seat at the table because of it.
The problem? AI adoption is pulling everyone back to level one. Instead of asking “what outcome are we trying to drive,” teams are asking “what can we do with AI.” That is a step backward, not forward. And it is happening at organizations of every size.
The Copernican shift
Sequeda used an analogy that applies directly to marketing leadership. Before Copernicus, everyone assumed Earth was the center of the universe. After Copernicus, we understood that Earth orbits the sun, and that reframing changed everything.
The equivalent shift for data and marketing teams: stop treating your technology stack as the center of the universe. The center is the work that needs to get done. Everything else, data, AI, analytics, governance, content production, campaign infrastructure, orbits around that.
For CMOs this means a discipline of asking: what is the actual business problem we are solving? Not “how do we use AI” but “customer satisfaction is declining and we need to understand why and fix it.” Once you define that, then you figure out where AI accelerates the path.
As Sequeda put it: “The way you’re going to do something with AI is by doing something with solving that work problem you’re trying to go do.”
Where marketing gets stuck
I see this play out in a specific pattern with mid-market marketing teams. There is board pressure to “do AI.” The CMO responds by adding AI tools to the stack, maybe an AI content tool, an AI ad optimization layer, an AI reporting dashboard. Each tool has a vendor demo that looks impressive.
Six months later, someone asks what changed. The answer is often: the inputs changed, but the outcomes did not.
This is the hammer looking for a nail problem. You have a tool and you are reverse-engineering a use case for it. That is not strategy.
What better decisions actually look like
If you want to make better AI decisions as a CMO, here is the sequence that holds up:
Start with one specific question the business cannot answer well right now. Not a category of questions, not a strategy area, one question. Something like: why are we losing enterprise deals in the final stage of the pipeline? What is driving the drop in repeat purchase rate among our best customers?
Understand why answering that question is painful today. Who is asking it, what they are doing to answer it now, where the data gaps are, and what it costs the business when the answer is wrong or slow.
Only then introduce AI into the conversation. Not as a starting point but as an accelerant to answering the question you have already defined.
This sounds simple. It is not easy, because the pressure to move fast on AI is real and it comes from above. But the teams I work with who are driving actual business results are the ones who kept the outcome at the center and treated AI as one of several means to get there.
The feedback loop that actually matters
One more thing Sequeda raised that deserves attention for CMOs specifically: the gap between insight and action.
Marketing teams generate insights. They build reports. They present findings. But too often the loop between “here is what the data says” and “here is the action the business took as a result” is weak or nonexistent.
AI can help close that loop, but only if you design for it. If your AI initiative produces a smarter dashboard that still sits in a slide deck that feeds a meeting that produces a follow-up meeting, you have not changed anything meaningful. The question to ask is not “how do we get better insights with AI” but “how does this insight connect to a decision and then to an action.”
That is the standard worth holding yourself to.
Find out how to grow your marketing with the professional marketing minds at Ambient Array.
Original source: Juan Sequeda and the Data Over Cocktails Podcast