AI Leadership Needs More Than Tools: Why Change Management, Governance, and Human-Centered Strategy Matter

We are not in a period of AI transformation. We are entering the age of Humanity.

The conversation keeps getting louder. AI this. AI that. Every board wants an ROI slide. Every function leader is running a pilot. And somewhere in the middle of all of it, a lot of smart, experienced executives are quietly asking themselves: Am I leading this, or am I being dragged into it?

At a recent Wednesday Women virtual event, three executive women sat down to talk about what it actually feels like to shepherd AI transformation from the inside — not from a demo stage, not from a venture capital pitch, but from the real seats where decisions land.

Megan Barbier, CHRO at Xactly Corp, is leading AI integration from the people seat. MJ Patent, CMO at Logically, is navigating the pressure cooker of revenue-driven AI expectations for her - and their customers. And Sophia Kindle, a strategic operations executive who oversaw AI transformation at Boeing, has seen what happens when the change management plan is an afterthought.

WATCH THE REPLAY

Facilitated by Cofounder of Wednesday Women, Melissa Moody, this was of the most grounded, honest conversations about executive AI leadership you'll hear. No sponsor-led, visionary fluff. But real talk with real humans thinking about hard-to-answer questions. Here are some of the session's key takeaways — and what it means for you and your business.

The Moment It Gets Real

For MJ, the turning point wasn't a headline or a board presentation. It was chaos.

"Everyone was off doing their own platforms and tools. There was no coordinated effort. How are we going to measure impact? How are we going to share this across teams?" she said. "It can turn into the wild west very quickly."

That's the moment a lot of organizations are in right now — and it's not a failure. It's a signal. Chaos is the natural outcome of enthusiasm without structure, and enthusiasm without structure is what happens when leaders endorse AI but don't step in to actually lead it.

Megan draws a sharp line between the two. Endorsing AI looks like saying yes to the provisioning request. Leading through AI looks like asking a harder question: What's on the other side of this?

"This is not a new tool being deployed. This is actually a transformation. Leaders that are leaning in are re-envisioning the future — what can this be, how is this maybe changing work as we know it and how we connect with each other?"

Those two postures — deployer versus re-envisioner — produce completely different outcomes. One creates a patchwork of tools and a mounting shadow AI problem. The other creates a foundation organizations can actually build on.

Change Management Isn't the Follow-Up. It's the Starting Line.

Sophia has been leading transformations for 15 years. She's watched change management get introduced at the testing phase, after decisions have been made and momentum is already baked in. With most technology shifts, that's bad practice. With AI, it's a near-guarantee of failure.

"If you don't have change management at the very beginning, it will fail," she said. "And we're seeing that from a lot of companies."

What makes AI different from every other enterprise technology rollout is scope. When organizations moved to the cloud, the impact was concentrated — developers, application managers, IT. It was significant, but it was bounded. AI is not bounded. It touches every employee, every role, every workflow. The learning and development function doesn't need to be updated; it needs to be redesigned from the ground up.

"We want to reskill people to think about their job differently," Sophia said. "If now I'm in a customer service or help desk role that's being managed by AI, I need to become an AI engineer — someone who creates agents that know how to do the work as environments change. That becomes your value."

This is not a small shift. It requires organizations to ask a question most have never formally asked: Who do we want our people to be on the other side of this?

That question is uncomfortable. However, it also might be the most important strategic question on the table right now.

The Speed Trap — and the Courage It Takes to Name It

There's a pressure pattern showing up across executive functions, and MJ named it directly: boards and shareholders pushing for ROI, asking why teams aren't moving faster, demanding AI results that are visible on a balance sheet. CMOs and CROs are feeling it most acutely, she said, because the applications to revenue-driving functions are obvious and immediate. But the infrastructure — governance, policy, data security, IT support — isn't keeping pace.

"IT is not equipped to help functional leaders through the change management," she said. "They're not equipped to develop policy, procedures, or governance. And so you have functional leaders running after it and implementing everything as quickly as possible for the board's sake."

Megan puts the cost of that approach clearly: rushing through AI deployment the way companies rushed to the cloud — hoping for savings, discovering complexity, and then quietly absorbing the cost of rollbacks and rehiring.

Cofounder of Wednesday Women and event host Melissa Moody cited a striking statistic shared in the event chat: 62% of organizations had already deployed live AI agents in customer-facing functions. Of those, 74% had already rolled them back. Among the most governance-mature organizations, that number was 84%.

Every rollback has a price. Every "we cut 200 people and then had to rehire" story has a price. Those prices are measurable — they just tend to appear on different spreadsheets than the ones executives showed the board when they made the original case.

"I'm constantly crusading against this measure of AI outcomes," Megan said. "I actually think it's going to get more expensive before it gets less expensive."

That's not a popular thing to say. It also happens to be true, and leaders willing to say it — clearly, calmly, with data — are doing their organizations a genuine service.

Who Owns the Time AI Saves?

Here's the question that had every participant thinking deeply, our event guests divided in their poll response, and one that most organizations haven't even begun to think about.

If an employee uses AI to complete in two hours what used to take eight — who owns the remaining six hours?

The instinctive corporate answer is: the company does. You're paid for full-time work. But Megan, coming from a CHRO perspective and a performance-outcomes culture, isn't so sure. "We've been pushing hard on outcomes over output," she said. "Are we being disingenuous if we now say, the time is ours?"

MJ took it further. What if AI capability became a compensation lever? What if reaching your performance outcomes unlocked flexibility — time, latitude, even the ability to take on other work? "Maybe we don't pay salaries anymore. Maybe we pay on outcomes."

This is not a settled question. It's not even a common one yet. But it will be. And organizations that start having this conversation now — before they're reacting to employee expectations they didn't anticipate — will be in a very different position than those that don't.

Human in the Loop Is Good. Human on the Loop Is Better.

Sophia is unequivocal about one thing: AI agents cannot carry accountability. They can't replicate human connection. They won't be held responsible when something goes wrong, and no board is going to accept "the AI agent decided that" as an explanation. The human has to remain in the loop — especially in high-stakes, high-visibility functions.

Megan took this further and re-framed the point in another, future-thinking way. The goal, she argues, shouldn't be to keep humans managing AI output. It should be to elevate what humans are doing entirely.

"I would love for us to be graduating from human in the loop to human on the loop," she said. "The agents are doing the work we were never meant to do — the reconciliation, the double-checking, the menial, the tactical. We're actually getting called to a higher purpose, which is the expression of our humanity."

That framing — AI handling the work humans were never meant to do — reorients the whole conversation. The question isn't "which jobs will AI replace?" It's "what are humans actually here to do, and how do we build systems that let them do more of it?"

The leaders who figure that out first won't just survive the transition. They'll define what work looks like on the other side.

The Loud, Uncomfortable Truth About Those Layoffs

Several major companies have publicly framed large-scale layoffs as AI-driven efficiency decisions. Sophia, Megan, and MJ didn't let that framing go unchallenged.

"They had non-realistic expectations of what AI was capable of," MJ said. "Applying AI into a process doesn't mean it takes two seconds and it's perfect. Instead of taking months, it might take a few weeks — but there's still a few weeks. There's still a human component you have to build into that process."

The math that made those layoffs look good on paper didn't survive contact with reality. Companies that cut deeply had to rehire. The recruiting and training costs of the return wiped out — and in many cases exceeded — the savings they'd claimed. And then, as Sophia pointed out, there's a macro layer that rarely makes it into the efficiency calculation: the people who lost those jobs were also the people buying products. Layoffs at scale affect consumer demand. That's not a soft concern. That's an income statement.

Host Melissa noted that Jensen Huang of Nvidia recently called leaders who brand layoffs as AI savings "cowardly." It's a harsh word. Megan added context worth keeping: some of it is also just branding. "Those pressures from the board — I need to react, I need to do something — and I might as well brand it as AI because it's believable."

Real leadership, in this moment, looks like controlling the pace instead of just responding to it.

Enabling Your People: What's Actually Working

The conversation turned practical toward the end, and a few concrete approaches stood out.

• First-line managers as AI champions. Sophia was clear: the most critical layer in any AI rollout is not the C-suite — it's the people closest to the individual contributors. "If they are your champions, they will unleash that freedom and that exploratory spirit on their teams." Senior leadership sets the vision; managers make it real.

  
  

• Gamification. Recognizing employees who bring AI tips back to team meetings, building channels where people can share what they've discovered, making the learning visible and social — these create momentum in ways that mandatory training rarely does.

  
  

• Showing the art of the possible through unlikely people. Megan made a point worth underlining: the biggest leap forward often comes when someone who's not expected to be technical shows what they can do. "When our office manager rolled out something amazing — no JSON, no Python — well, so can I. It takes down this idea of I'm not technical enough."

  
  

• Letting personal learning happen. For teams waiting on IT to provision access, Megan's approach is practical: tell your people to build skills on their own time, on personal instances, without connecting to company data. When their seat opens up, they're already a lap ahead.

What Kind of Leader Does This Moment Need?

At the close of the conversation, Megan offered something that was less a strategy and more a disposition.

"Trust and vulnerability are going to be invaluable," she said. "We're all learning together. Let people see you figuring it out in real time."

That's a departure from the traditional model of the executive as the person in the room with the answers. This transformation doesn't work that way. Nobody has the answers. The organizations that will come out ahead are the ones where leaders are honest about that — where showing uncertainty is a feature of the culture, not a liability.

There's also a harder prediction embedded in what she said: the spread is going to widen. Fast. Organizations ahead of the curve will pull further ahead. Those behind will fall further behind. And the window to close that gap is narrower than most people want to believe.

The Reframe That Changes Everything

Megan said something in passing that deserves to not be in passing.

"Are we really in an AI transformation? Or are we actually in the age of humanity?"

The more you sit with that question, the more it reshapes the whole conversation. The technology is moving too fast to manage. The governance is too slow to catch up. The ROI math keeps not working the way it was supposed to.

What doesn't become obsolete is the human capacity for judgment, connection, accountability, and creativity. What's being demanded of leaders right now — more than any technical literacy, any AI fluency, any prompt engineering skill — is the willingness to be fully human in a moment when everything around them is trying to automate.

That's not a consolation prize. It's the job.

MJ Patent, Megan Barbier, and Sophia Kindle are executive members of Wednesday Women. Follow their posts here on LinkedIn. MJ's company, Logically, has also made available a free workbook of AI implementation templates from their Logicon conference on this topic.

Wednesday Women hosts virtual conversations for executive women across industries. To learn more about the movement, the membership, or hear about upcoming events, subscribe here.