Back to Blogs
Drew Neisser is the founder of CMO Huddles and a globally recognized authority on B2B marketing. He’s an AdAge columnist, LinkedIn TopVoice, leading CMO coach, podcast host & friend of penguins everywhere.
CMOs need firsthand AI experience without sacrificing customer understanding or team development. Drew Neisser and four-time CMO Eric Eden explore how marketing leaders can balance technical fluency, systems thinking, coaching, and business strategy. Their conversation offers practical questions for protecting leadership time, evaluating AI costs, and connecting experimentation to meaningful business outcomes rather than activity.
Every hour a CMO spends learning an AI tool has an opportunity cost. That hour might produce a breakthrough, expose an expensive limitation, or make the next conversation with a vendor considerably more productive. It might also replace a customer conversation, a coaching session, or the meeting that finally gets sales and marketing moving in the same direction.
That trade-off became the heart of a recent conversation with Eric Eden of Thinking Deeply, a marketing leader who has held the CMO role at least four times. Eric and I are enthusiastic AI users, and neither of us thinks CMOs can afford to remain spectators. We disagree about how much personal experimentation leaders need and how to keep that learning from consuming the job.
For CMOs navigating this transition, our conversation offers a useful tension to hold onto: develop enough firsthand understanding to make informed decisions while protecting the customer relationships and team capabilities those decisions are supposed to serve.
My starting point is that great CMOs are great business leaders. They pick the team, set the direction, allocate resources, and build the relationships that allow marketing to succeed. Customer understanding informs all of those responsibilities.
I’ve long argued that CMOs should spend roughly 25% of their time with customers, partners, and forward-thinking vendors. That is my recommendation, not a research-derived benchmark. Those conversations help leaders understand changing buying behavior, recognize emerging needs, and distinguish a genuine opportunity from an impressive demonstration.
During our conversation, I pushed Eric on the opportunity cost of deeper technical involvement. A CMO who spends their time “working the relationships across the organization and talking to customers,” I argued, has an advantage over one who becomes absorbed in designing the system. Knowing how to produce something faster does not automatically tell you whether customers need it.
Your customers are not a distraction from your AI transformation.
Eric’s response sharpened the discussion. He wasn’t suggesting CMOs abandon customer conversations to become system administrators. He was arguing that leaders need enough direct experience to recognize when their teams are working within constraints that technology has already changed.
“You only learn by rolling up your sleeves and doing some of this stuff,” Eric told me. His next sentence matters just as much: “And once you’ve experimented, you don’t have to run the system.” For him, experimentation builds judgment that a leader can apply long after someone else takes responsibility for execution.
Eric described website projects where he believes AI-enabled workflows can substantially compress timelines. His point was that a CMO who has seen those workflows in action can challenge an estimate, investigate alternatives, and ask more useful questions. The actual timeline still depends on the project’s scope, integrations, review requirements, and content readiness.
That is a persuasive argument. A leader who knows only the previous generation of tools may accept the previous generation of assumptions, including what work costs and how long it takes. Personal experimentation can reveal possibilities that never appear in a status report.
I’ve experienced that myself. Building our AI Maturity Calculator and working on its landing page showed me that I could create useful things without writing the underlying code. I also described an agent I tried to build that consumed enough resources to make me turn it off.
Both experiences were educational. One expanded my sense of what was possible; the other reminded me that enthusiasm is a poor substitute for understanding operating costs.
The leadership challenge becomes especially clear when a new capability collides with an established team. Eric described a client meeting where marketers raised concerns about connecting a new assessment tool to their CRM. He believed the connection was straightforward, but the conversation escalated when the CEO challenged the team’s objections.
“That’s not what I wanted to happen,” Eric said. He had intended to help the team move forward. Instead, the demonstration became a source of friction, illustrating how quickly a discussion about technical feasibility can turn into a judgment about people.
For me, that story raises questions every CMO should consider. Does the team lack knowledge, confidence, capacity, or permission? Is someone raising a legitimate implementation concern, or relying on an assumption that deserves another look?
Those situations require curiosity from the leader. Ask someone to demonstrate the proposed approach, surface the dependencies, and explain what still needs human review. Give the team room to learn before treating every hesitation as resistance.
Eric’s experience adds an important counterweight: leaders also need to recognize when familiar processes are being defended without sufficient examination. Developing people includes helping them let go of work that no longer needs to be done the same way.
Eventually, Eric clarified the phrase at the center of our disagreement. “It’s not that you need to be like coding,” he said. “I’m saying people have to be more systems thinkers.” That moved our conversation toward how work connects across people, tools, and departments.
One question particularly interests him: “How to do multiplayer AI, how teams can use AI together instead of operating in silos.” An individual productivity gain becomes more consequential when the resulting work can be shared, reviewed, maintained, and used by others. Those connections require deliberate choices.
Eric pointed to AI-enabled website tools that send information into a CRM or help book meetings for sales. Marketing needs to understand what happens after the interaction, and sales needs to understand what it is receiving. Faster production at one stage can create confusion downstream if the teams have different expectations.
This is where technical fluency and relationship-building reinforce each other. The CMO needs enough understanding to ask about the handoffs and enough organizational credibility to get the right people working through them. Strong marketing operations and technical specialists remain essential, with the leader responsible for keeping the work connected to the business objective.
Eric’s most concrete examples involved the economics of individual tasks. He described spending approximately $100 in tokens on a market-sizing research project and about $25 on a presentation. Those were his reported costs for specific assignments, not universal benchmarks.
“I look at what each job actually costs across the different tools,” he explained. From there, he considers whether another approach could deliver comparable quality more economically, or whether the importance of the assignment justifies the expense. His advice was straightforward: “Be a good steward of capital.”
I agree, and I want CMOs to carry that thinking into budget allocation. AI spending competes with people, existing technology, and programs. A collection of inexpensive experiments can still become a meaningful expense, particularly when maintenance, review, and integration work accumulate.
In our conversation, I suggested starting with a defined growth opportunity and building the economics around it. What would entering a new market require? Where could AI help, what would it cost, and what evidence would justify further investment?
The token bill belongs in the business case.
This approach also gives the team a clearer assignment. People can evaluate whether a workflow advances an agreed objective rather than trying to demonstrate AI enthusiasm through sheer activity.
Eric and I did not arrive at a precise formula for dividing a CMO’s week. His experience leads him to place greater emphasis on learning by doing; mine makes me protective of the time leaders spend understanding customers and developing people. Both concerns deserve a place on the calendar.
My practical threshold is regular personal AI use, experience building at least one meaningful workflow or agent, and a working understanding of costs and limitations. From there, the question becomes what additional learning will materially improve your decisions. A focused experiment can be valuable without becoming a permanent second job.
For your next calendar review, consider three questions:
The balance will vary by company, team, and stage of adoption. It should also change as your capabilities grow. What deserves ongoing scrutiny is whether your time is improving the decisions, relationships, and people on which marketing performance depends.
Eric and I will both be at the CMO Super Huddle, October 22–23 in Palo Alto, where this conversation will continue. Join us to compare experiences with other marketing leaders, and bring your own answer: how are you making room for AI fluency while protecting the work only you can do as CMO?
Pick a recurring task or decision tied to a business priority, define what a good result looks like, and set a time and spending limit. Review the output and the complete workflow with a relevant specialist before expanding its use.
Specialists can own implementation, integrations, maintenance, and technical troubleshooting. The CMO remains responsible for business priorities, resource allocation, acceptable outcomes, and making sure the right people review risks and performance.
Ask the team to identify the specific dependency, risk, or resource gap and demonstrate it where possible. A focused test with an appropriate specialist can help resolve uncertainty without dismissing the team's expertise or accepting an outdated assumption.
No. Drew recommends spending roughly 25% of leadership time with customers, partners, and forward-thinking vendors. It is his practical recommendation, not a research-derived benchmark, and leaders should adapt it to their organization and responsibilities.