He didn’t push back.

“Sometimes,” he said, “different parts of the org may be less willing to change, maybe to your point, because it’s their job not to do that.”

It was an honest moment in a 40-minute conversation with the executive overseeing half of LinkedIn’s workforce—and the one his company’s own research, released this week, most clearly points toward.

The survey of 1,252 C-suite leaders in the U.S., U.K., and India finds half of executives acknowledge they don’t have clear visibility into the roles and skills their organizations will need as AI matures—what LinkedIn’s report calls a “workforce blind spot.” Seventy-eight percent say they are moving faster on AI than they can effectively measure.

Those numbers are striking. What Lobosco adds to them, from his weekly conversations with CMOs, CHROs, and CROs, is a more uncomfortable layer: The blind spot isn’t just about uncertainty. It’s about structure.

“We’re at this place where the C-suite is navigating this moment without a real playbook they can rely on,” he told Fortune. That’s pretty well known by now, but what the survey quietly surfaces is the standard response to that problem, the top-down mandate to transform, may be making things worse.

This moment really calls for good “change management,” Lobosco said, and that “doesn’t work if it’s top down. It doesn’t work if you’re not bringing everyone along on the journey.” Transformation succeeds, in his framework, when employees see AI as a “career accelerant” rather than a threat—and that perception, he said, can’t be decreed from the executive suite.

But he was equally clear on the other side of the paradox: Executives can’t delegate it either.

“You can’t ask your team to do something that you don’t know how to do yourself,” he said. “You can’t not be very proficient in using these tools. People will sniff it out immediately.” The companies managing this best, in his observation, are the ones where the C-suite leads from the front—proficient users of the tools, not just advocates for the strategy.

That leaves a narrow window. Too top-down and it fails. Too delegated and it fails. Success requires leaders who are genuinely fluent, empowering teams from a position of real knowledge, with clear end goals in sight beyond just “adopt AI.” By his own account, most companies are not there yet.

The research backs him up. Eighty-two percent of C-suite leaders say entirely new AI-related roles have grown inside their organizations since 2022—Lobosco described role titles like forward-deployed engineers, AI engineers, responsible AI architects, go-to-market engineers. But the same leaders creating those roles can’t describe what the workforce around them will look like in two years. They’re redesigning the plane while flying it, without, to paraphrase Lobosco, a flying manual.

His internal solution at LinkedIn is deliberately bounded. Rather than trying to turn everyone in a large commercial organization into an AI power user, he’s focused on building what he calls a “cockpit”—infrastructure that removes friction, automates preparatory work, and frees his team to spend 90% of their time with customers.

“I don’t think there was ever a goal of turning everyone into cyborgs,” he said, referring to another model Fortune has reported on. The goal is to unlock more time for the judgment, creativity, and relationship work that customers actually need.

It’s a more honest frame than the generic “AI fluency for all” push. It also quietly concedes that the people who most need to change—senior executives who built careers executing reliable playbooks—may be the hardest to reach.

That concession has backing. Carolyn Dewar, who founded McKinsey’s CEO practice and has spent decades advising chief executives, wrote in a Fortune column this month the execution discipline markets rewarded for the last decade is now “a liability.” The environment has shifted faster than leadership teams can adapt, she argued.

“You cannot OKR your way out of a moment that requires fundamental strategic rethinking,” she wrote, referring to objectives and key results-thinking.

The capabilities now in demand, according to Dewar, are judgment, imagination, strategic courage under uncertainty, “deeply human leadership capabilities.” That’s nearly identical to what Lobosco said he hears at the top of every CXO call.

At the Fortune COO Summit in Scottsdale last month, Okta President and COO Eric Kelleher offered the operational view of the same problem. The hardest part of AI transformation isn’t the technology, he told a room of senior operations executives—it’s the managers.

“We have trained every manager in the world to think about one thing and that is: What’s their headcount,” Kelleher said. The shift to thinking about “human workers and digital workers” as distinct categories of labor—budgeting for both, designing org charts around both—is “a much harder problem than getting people to experiment with Claude Code.” Companies, his diagnosis ran, are in collective denial about redesigning work itself.

LinkedIn’s broader labor market data shows global hiring is still 20% to 30% below pre-pandemic levels. Asking executives to learn in public, improvise with unproven tools, and redesign organizational structures simultaneously, while meeting the quarter, is a lot.

Lobosco’s read is most companies are still in the early stages of what will ultimately be a long redesign—the electricity problem. Factories bolted electricity onto steam engines for years before anyone thought to redesign the factory floor. The insight wasn’t wrong; the timeline was. LinkedIn CEO Ryan Roslansky has reportedly made this a topic of internal keynotes.

What LinkedIn’s survey doesn’t capture—and what the conversation with Lobosco does—is how rational the resistance actually is. If you built your career on executing a reliable playbook, the person asking you to abandon it for blank-sheet thinking has a credibility problem. The blind spot isn’t simply that leaders don’t know what they need. It’s that admitting it, publicly, to their teams, is its own kind of risk.

Lobosco isn’t naive about this. He frames his own approach around leading from the front precisely because executives who can’t show their work have no standing to ask others to change. Whether the rest of the C-suite follows that logic is the question the survey can’t answer.

“Companies are still making moves,” he said. “But they’re still not exactly sure where this ends.”