9 Reasons AI Adoption Stalls in B2B Teams

Quick summary: Nine specific reasons AI initiatives stall in B2B marketing teams, from the wrong measurement metrics to the middle manager quietly sinking every idea in the room. Each one is fixable. None of them are about the tools.

---

“Everybody has an AI mandate,” Sandra Miley told a room of CMOs at a recent CMO Huddles Strategy Lab. “Not everyone has an AI strategy. There may be a correlation here.”

Miley works with enterprises on the full arc of AI transformation, from initial skills training through multi-agent workflow deployment. She’s seen what accelerates adoption and what kills it. The pattern she's noticed is consistent: The technology is rarely the problem. The people, the culture, and the change management almost always are.

Here are nine reasons AI initiatives stall, drawn from her sessions in our San Francisco and Palo Alto Strategy Labs.

1. You Have a Mandate, Not a Strategy

An AI mandate says: We are doing AI. An AI strategy says: We are using AI to accomplish this specific thing, and here is how we will know it is working.

The teams Miley sees making real progress have done the work of connecting AI to business outcomes. The teams that are spinning have an executive directive, a collection of tools, and not much else in between. If your team cannot answer which problem the AI is solving and how it connects to a company priority, you do not yet have a strategy, you have activity.

2. You Told Your Team to ‘Go Play’

Telling your marketing team to "go play" sounds empowering. In practice, most people in organizations are not natural experimenters. They want to know what to do, why to do it, and what success looks like. When leaders say "Go figure it out," the majority of the team tries something once, gets a mediocre result, decides AI is not for them, and quietly moves on.

“Most people I found in corporations want binary answers,” Miley says. “If it doesn’t work the first time, they will reject it entirely. ‘I got a bad answer from ChatGPT, therefore I do not use ChatGPT.’ That’s how the majority of people work. And we have a huge blind spot about this in Silicon Valley.”

The fix is specificity. Don't say, "Go play with AI." Instead, try, "Go use AI to do this specific task, in this specific way, and here is what good looks like."

3. You’re Training On Tools, Not Tasks

Generic AI training produces generic AI adoption. Showing your team a demo of what ChatGPT can do in the abstract does not change how anyone works on Monday morning.

Miley’s recommendation is to train on tasks, not tools. Here are some examples she suggested:

• Show your demand gen manager how to build an automation for their specific lead workflow
• Show your content marketer how to systematize their brief-to-draft process
• Show your product marketer how to build a GPT loaded with competitive intelligence

When people see AI eliminate something from their actual job, resistance drops faster than any awareness session can achieve.

4. You Didn't Start with What People Hate Doing

One of the most effective techniques for breaking resistance is to ask your team what they dread doing most, and start there. One CMO shared how her team used AI to shorten the time it took to create a Monday stand-up report from hours to minutes. Another tackled the RFP response process that was consuming his team’s time every week.

Miley endorsed this approach directly: When people see AI eliminate something they genuinely dread, they become curious about what else it can do. The drudge work is your fastest path to a believer.

5. Your Middle Managers Are Quietly Sinking It

Every CMO in Miley's sessions had encountered resistance to AI adoption. If your AI initiative keeps stalling in ways you cannot quite explain, look at the middle layer first.

Middle managers have built their authority on owning workflows and controlling information flows. When you redesign a workflow to be AI-assisted, you are not just making it faster. You are making it more transparent, more accessible, and less dependent on any single person. That is a direct threat to how they exercise power.

“They will not be openly resisting you in the meeting,” Miley says. “It’s very subtle. And the only thing you know when you come away from it is: Everything just got sunk somehow. Every idea. I couldn’t put it together.”

6. You’re Not Recording the Meetings

Miley found a technique that changed her entire approach to launching an AI initiative: She took the meeting transcript, put it into Claude, and asked it to analyze what had happened.

The result stopped her cold. Claude came back and said, in effect: Stop. You cannot move forward with technology until you address the human problem. There is a change management problem with this individual. They are complexifying, obfuscating, redirecting.

“What you’re going to find,” she says, “is that most of us remember the last thing said and the feeling when we walked out of a meeting. We seldom remember the first things, the middle things. When AI analyzes the full transcript, it will recognize the pattern.”

Record every meeting where your AI initiative is on the agenda. When something feels off, run the transcript through Claude and ask it to identify the agendas and surface the patterns. It will name the problem you could not quite articulate.

7. You Haven’t Found Your Champions Yet

Every organization has people who will run with AI before you've finished explaining it. Baseline training reveals who they are. They come back with use cases you did not demonstrate. They start asking about the next stage before they have been introduced to the current one.

Miley is direct about what to do when you find them: Give them a role. Give them a voice. Put them in front of their peers. These champions will do more for your team’s adoption than any external resource, because peer-to-peer proof is what makes it safe for everyone else to try.

⚠️  Do not wait until your initiative has momentum to identify your champions. Find them in the first training session and activate them immediately.

8. You Went National Before Local

Miley’s rule for pilots, especially agentic ones: Crawl, walk, run. Start with one use case, one workflow, one team. Prove the ROI in a small, controlled way, then expand.

The failure mode she sees repeatedly is when organizations announce a company-wide AI initiative, roll it out broadly before the governance or training is in place, watch it produce inconsistent results, and conclude that AI does not work for them. The problem was not the technology. It was the sequencing.

Sandbox it first. One customer, one briefing, one workflow. Get it right at that scale before you blow it out.

⚠️  When working with agents specifically: Agents move fast and can take things off the rails just as fast, so test in a sandboxed environment before connecting to live systems. Make sure someone on the team understands every data source the agent touches and every action it could take.

9. You’re measuring the wrong things

Hours spent in AI tools is not a metric. Token consumption leaderboards measure activity, not outcomes. Someone burning two million tokens a day might be running in circles. Someone spending thirty minutes a day might be running a workflow that is doing the work of three people. You cannot tell from the numbers alone.

Miley ties AI ROI to executive-level KPIs. The CEO wants to know what the organization can now do that it could not do before. The CFO wants time-to-value and forecasted capacity. The CHRO wants to know whether employee satisfaction is higher among teams using AI tools.

“Tie your metrics to executive KPIs and OKRs,” Miley says. “The outcome. The sale. The revenue. The employee satisfaction. The capacity increase. Those are the things to focus on. ‘Hours spent in ChatGPT’ is not a metric.”

One concrete benchmark Miley points to is that when teams roll out and train properly on AI tools, a 12 to 25 percent capacity increase without additional headcount is a realistic early target. That is the number worth tracking, not the leaderboard.

The Through Line

Miley closes every session with the same observation: AI transformation is highly humanistic. The tools are almost never the problem; the culture, change management, and clarity of leadership narrative are where initiatives succeed or fail.

The CMOs getting this right are not just running more efficient teams. They have teams that feel genuinely equipped for where the market is going, building capabilities their competitors have not figured out yet, and spending more time on the work that actually moves companies forward.

That outcome does not start with the tools, it starts with an honest answer to the question Sandra Miley asks at the start of every engagement: Describe your team’s AI maturity in one word. And then be honest about what that word actually means.

---

Sandra Miley is an AI transformation strategist who works with enterprises on the full arc of AI adoption, from initial skills training through multi-agent workflow deployment.

Want more?

For a framework on where your team sits on the AI maturity curve, read: AI Maturity Model for Marketing Teams: The 6 Stages You Can’t Skip.

For the four-part AI adoption strategy that drives behavior change, read: Your Team Is Doing ‘Random Acts of AI.’ Here’s How to Fix It.

It’s warmer in the huddle. CMO Huddles brings together senior B2B marketing leaders for candid, peer-to-peer conversations on the challenges that matter most. Starter membership is free.

FAQs on AI Adoption