Back to Newsletters
May’s Leader Huddles showed a clear shift in the AI conversation across the CMO colony. CMOs are no longer asking whether they should use AI. They are asking how to turn scattered experiments into a better way to run marketing. The sharper conversations are moving beyond productivity wins into measurement, build-versus-buy decisions, agent guardrails, sales enablement, team readiness, and operating model design. The opportunity is not simply to create more content or automate more tasks. It is to get the whole huddle moving with a marketing organization that can use AI repeatedly, responsibly, and measurably.
AI is making it easier to launch more content, campaigns, and workflows than ever before. But several CMOs admitted that measurement is not keeping up. One Huddler put it plainly: “Too much is being launched in real time, and we can’t catch up with the measurement as fast.”
Another noted that the team can point to individual productivity wins, but not yet “the total impact to a function or a team.” That is the trap. Faster output is only useful if it leads to better business outcomes or faster learning. Otherwise, marketing may simply create more noise at a higher velocity.
The practical win is to pick one metric the CFO already cares about. One CMO chose lower cost per acquisition because it factors in people, variable costs, and business impact. Another framed the AI opportunity as scaling from $40 million to $100 million “without scaling my BDR team.”
That is the kind of AI story boards understand. Not just “we saved six hours,” but “we grew without adding expense.” CMOs should make sure AI success does not live only in productivity anecdotes. It needs to connect to a business number that already matters.
A few months ago, many teams wanted to build their own agents. Now CMOs are seeing the fine print. Custom workflows can break. Tools can change. Token costs can add up. And when an agent gets one small thing wrong, the team may not know how to fix it.
As one Huddler warned, “If you build it, and it breaks even in its smallest way, no one knows how to fix it.” That does not mean teams should never build. It means they should build where differentiation matters and buy where reliability matters.
One CMO shared a useful example from an AI SDR rollout. “We had already signed a contract with Qualified,” they said, before seeing 1Mind at a CMO Huddles Strategy Lab. While 1Mind “blew me away,” the Qualified implementation was already producing results: 20 meetings booked and 12 opportunities worth $29 million.
Even better, the AI SDR was uncovering opportunities “at all hours of the day and night,” booking meetings while human SDRs were sleeping. That is not AI as a toy. That is AI as leverage.
One of the most useful cautionary stories came from a CMO describing an AI workflow that could build a sophisticated ABM strategy but still got a calendar date wrong. “All you have to do is let the agent run,” they said, “and then sometimes it can’t even get a date right.”
That tiny mistake matters. As another CMO observed, one obvious flaw can cause the recipient to dismiss everything else that is good. The more mature answer is not to trust agents everywhere. It is to decide which agents can act alone, which can recommend, and which need a human review before anything customer-facing goes live.
Some CMOs are already building narrow, useful “brains” for specific jobs: Event-vetting agents, citation-monitoring agents, persona-checking GPTs, and sales enablement brains that draw from approved thought leadership.
That last use case matters. One CMO described the dream scenario of sales teams using marketing content inside AI prompts and getting stronger output because the thought leadership added proof points and case studies. For years, marketers have wondered whether sales was using their content. AI may finally make that content easier to find, trust, and apply.
Many CMOs have AI task forces. Fewer have true AI operating models. One CMO described the tension perfectly: “It still feels like it’s off the side of a number of people’s desks.” Another said they had plenty of AI “superstars,” but not a clear way to turn ideas into scalable workflows.
That is the next maturity gap. A task force can spark motion, but transformation requires ownership, skills, governance, and repeatable ways to turn ideas into working systems.
Not everyone on the marketing team needs to become a technical wizard. But every team needs a way to pair business imagination with AI-building capability. In the Huddles, this sounded like a buddy system: One AI power user working with a small group of marketers to test, build, and refine workflows.
Human creativity plus technical fluency is how experimentation becomes capability. Capability is how AI becomes an operating model.
At the Imaginarium Summit, Noah Brier, founder of Alephic, offered a useful warning: AI is “a mirror, not a crystal ball.” In other words, AI reflects the organization using it
If processes are messy, AI will expose that. If data is disconnected, AI will expose that. If teams are undertrained, AI will expose that too. That can feel uncomfortable, but it also gives CMOs a practical roadmap.
Before chasing every new AI possibility, leaders should ask what AI is revealing about the organization. Where are workflows unclear? Where is data weak? Where do teams lack confidence? Where does governance break down? Those answers are not side issues. They are the foundation of AI maturity.
The healthiest AI programs do not try to boil the ocean. They start with a real business problem, pick a meaningful metric, choose where to build and where to buy, and decide what level of human review each workflow requires.
That sequence creates focus. It also prevents the team from mistaking activity for transformation. The point is not to run more AI experiments. The point is to create repeatable ways for marketing to work better.
For CMOs, this is where leadership matters. AI operating models will not emerge from isolated enthusiasm alone. They need executive direction, practical guardrails, and a clear connection to business outcomes.
An AI operating model is the repeatable system a marketing organization uses to identify AI opportunities, prioritize use cases, assign ownership, apply guardrails, measure impact, and scale what works.
Productivity metrics are useful, but they rarely tell the whole story. A team may save time without improving pipeline, conversion, customer experience, or cost efficiency. CMOs need to connect AI work to business outcomes.
Build when the workflow creates meaningful differentiation or requires proprietary context. Buy when reliability, support, security, or speed matter more than uniqueness. The right answer will vary by use case.
CMOs should classify agents by risk. Some can run autonomously, some should recommend actions, and some need human approval before anything reaches a customer or prospect.
The next phase of AI in marketing is less about proving that AI can do interesting things and more about proving that the marketing organization can use AI well. Start with one real business problem. Pick one metric that matters. Choose where to build and where to buy. Put guardrails around agents. Pair curious marketers with technical builders. That is how teams move the whole colony from “look what AI can do” to “look what our marketing org can now become.”