Why Most AI Rollouts Stall and How CMOs Can Fix It
AI Coaching Demand: Nicole Leffer transitioned from Head of Marketing to a leading AI Advisor for marketing teams.
Adoption Barriers: Leffer identifies that access to AI tools doesn't equate to successful adoption; strategy is crucial.
Tool Focus: Leffer advises marketing teams to deeply invest in a single AI tool to enhance workflow efficiency.
Rollout Phases: Skipping AI adoption phases can lead to risks; Leffer emphasizes a structured approach.
Future Skills: AI offers potential for marketers, but future success relies on building relevant skills and insights.
Most CMOs didn't sign up to be their company's AI strategist. Sure, you committed to building brands, driving pipeline, and leading teams. Somewhere between 2022 and now, your job description quietly expanded.
Now, leaders are expected to evaluate every new AI tool, develop an adoption strategy, train the organization, and track what's changing.
On top of this, you're meant to do it all without a playbook, a dedicated budget, or any reduction in the original responsibilities. Nobody asked if this was reasonable, it simply become the expectation.
Want more from The CMO?
Sign up for a free membership to complete reading this article:
Have an account? Log In
AI Advisor, Nicole Leffer, has watched this play out across hundreds of marketing leaders, from Series A startups to Fortune 50 companies. And while she didn’t set out to become a sought-after CMO Advisor, her path became clear when she shared what she knew inside a CMO community in 2023, and more than a hundred marketing leaders signed up to hear her speak.
Since then, Leffer has been up-skilling B2B marketing teams on how to use AI, in practice, in workflows, and at scale.
Why Adoption Fails
Most of Leffer’s clients arrive with the same scenario, where they’ve rolled out an AI tool like ChatGPT, Gemini, Copilot, or Claude, and nothing has effectively changed.
I help move people from having the tool to actually building out sustainable, scalable workflows with the tool, moving to where its being put into place in a really great way.’
The issue, she says, isn't the tool. It's the assumption that access equals adoption. Except teams need a bridge from owning a subscription to building workflows that actually stick.
She underscores the importance of guidance as part of your rollout, and a degree of understanding that "you can't just give people a tool and expect they're just going to figure it out on their own.
When adoption is low, results are unclear, and leadership grows frustrated. The culprit is almost always the assumption that access equals adoption.
But teams need a bridge from owning a subscription to building workflows. Before deciding whether your AI investment is working, you need to define what it means when it does.
Start by gauging whether your team needs further training on how to use it. A license is not a strategy, and handing people a tool without that bridge is setting the investment up to fail quietly.
The Filter Problem
Marketing leaders are under a specific kind of pressure that shows up consistently across organizations. Now, you're expected to be the AI expert for the entire company. All while having no background in AI strategy and a day job that hasn’t slowed down. As Leffer explains, there's a reason why AI rollouts are stalling.
Leaders are being asked to become overnight experts in something that is a full-time job to be an expert in. I don’t even know any experts that actually feel like 100% up to date all the time.
The pace of change makes this worse. By the time a CMO has absorbed one development, the landscape has shifted again. What leaders actually need, Leffer argues, isn't more information, it's a filter.
She shares how the folks she works with are looking for guidance. "People just need somebody to tell them where to focus and what they can ignore. There's so much noise and so many people pretending they know what they're talking about."
The pace of change makes this worse.
By the time a CMO has absorbed one development, the landscape has shifted again.
Leffer is candid about how own information sources aren't the same ones she’d recommend to clients. Her role requires her to simultaneously track three very different horizons, a move that can be done in tandem if you're looking to set strategic milestones of your own.
From Leffer’s point of view, she has to meet clients in past, present, and future states. To do this, she asks, 'Where is their team today? And what is their team ready for?'
She emphasizes the importance of knowing where the tech is currently, focusing on "where is the tech going, so that we're preparing our skills six months, say a year from now." For cutting edge insights, Leffer looks to the folks on the frontlines of AI development.
I'm paying attention to AI engineers, the people who are making the tech, reading research papers, understanding the conversations that are happening behind the scenes at AI companies.
Now, the mistake most leaders make is trying to keep up with everything. You really need a filter to gauge what matters now, what can be safely deferred, and who you can trust help make these calls.
The outcome you want will come when you focus your efforts on what you can control. Ensure you know what matters now, what you can safely defer, and who you can trust to help you tell the difference.
Pick One Tool and Go Deep
This is the piece of Leffer's advice that tends to generate the most pushback, and that she stands behind most firmly. The single biggest operational mistake she sees marketing teams make is adopting too many AI tools at once.
If you have too many different tools, you don't know what you use for what and you don't actually know how to use any of them because they all change their interface every 15 minutes.
The platform-switching cycle is damaging. Teams hear that a competing tool is better for a specific use case. So, they migrate their workflows, rebuild their prompts, and then the original tool catches up, sometimes within days.
Plus, the cost of toggling between platforms adds up. Between a bigger AI bill, more cognitive load, and ironically, lower adoption, switching defeats any marginal performance difference.
Any benchmark comparison made even a month ago is likely obsolete. The performance differences between the major platforms are almost always smaller than the cost of switching.
Instead, a practical move looks like picking one of the four core tools. She mentions ChatGPT, Gemini, Copilot, or Claude, paying for it at the company level, and building portable workflows within.
Leffer shares that "even if you had done this testing, let's say just hypothetically ChatGPT versus Claude a month ago, that is 100% an out-of-date view today. It becomes 100% irrelevant. I have seen days where in the morning the opinion was obsolete by afternoon."
Before committing to one tool, by all means evaluate alternatives. But don’t migrate until the gap you see is significant and stable, and also applicable to how your team needs to use the tool.
Just pick one tool, learn it, build your workflows, build it portably. You’re so much better off than constantly trying to chase the shiny object.
Her advice isn’t new, but it’s relevant in a time when it’s all too easy to pop open another tab and you’re off to the races. Commit to making incremental progress, and start by choosing your tools wisely.
Don’t Skip the Sequence
Leffer maps AI rollout as a series of distinct phases. There’s basic chat interaction, prompt templates, custom GPTs and Gems, workflow automation, and finally agentic systems that can operate autonomously. This is a very high level view of what AI adoption might look like.
She shares how more organizations are taking shortcuts for fear of being outpaced by their competitors.
There are phases that every company goes through, and if you skip one of the steps you're not prepared for the next step. A lot of people are trying to skip the steps because they're afraid of falling behind.
The order matters. Teams that jump phases haven’t learned how to give AI the right level of instruction. And more importantly, they haven’t learned how to recognize when something has gone wrong. Leffer cautions operating without safeguards in place.
She continues, noting that risk of blindly trusting half-baked processes. "The problem could be significant. Like, maybe your AI just took your entire HubSpot database and published it on the internet."
She's also clear about where her own expertise ends. On questions of cybersecurity and technical risk, she flags the issue and directs clients to their tech teams. Plus, she’s noticing concerning behaviour from companies trying hard to be on the forefront.
They don't understand what they don't know and they're taking a lot of security risks. I think a lot of that is honestly going to come through marketing because people are recklessly connecting tools in ways that can create major risks, major problems.
Education comes before workflow-building, which comes before automation. Then agentic systems follow. A lot of organizations are trying to skip steps because they’re afraid of falling behind. And this fear is exactly what makes them vulnerable.
The risks at the agentic end are not theoretical. An undertrained team running autonomous AI workflows in a live environment is a genuine liability, not a competitive advantage.
What The Future Holds
When asked about what excites her most about AI right now, Leffer lights up. AI can support marketers in developing meaningful skills, learning what's possible, and finding ways to do work that once felt out of reach.
She's also spotting a quieter trend developing her client conversations, one that concerns her as much as the headline risks.
Some CMOs are openly anticipating a future where AI handles enough of the operational work that they no longer need to manage people at all.
She shares how in her conversations with marketing leaders, she hears things like, 'I can't wait until we get agentic and the AIs can do it and I don't have to deal with managing people.' And that, "those people weren't talking like that three years ago."
On the question of whether AI will eventually replace human creativity and taste, the jury is out. The most important thing you can do right now is build your team's skills. In part, to protect their relevance, but also yours.
Continue to invest in having a deep understanding of your customer. Finesse brand messaging and positioning so it's clear who you serve, what you offer, and why you're the best choice. In doing so, you'll be in a much stronger position than the folks who move the fastest—regardless of their tactics.
While the future may be uncertain, your leadership needs to guide teams through AI adoption and into what's next. AI transformation plans fail because teams never change how the work gets done. The CMOs seeing results today are the ones who’re moving past experimentation in an effort to build real operational muscle.
What’s Next?
Looking for more insights from experts? Join us at the The CMO Club. Learn from marketing leaders, and get the latest resources in your inbox.
