I once sat across a conference room table from a newly appointed CEO and told him not to believe me.
The situation was what you might politely call a standoff. An engineering manager with credentials, a team, and organizational authority on one side. Me on the other… one of many sales engineers, no direct reports, no title worth mentioning. What I did have was a reputation. When everything looked bleakest, when the customer relationship was on the edge, when the deal needed saving or the situation needed defusing… that’s when they’d send me in.
The engineering manager and I had fundamentally different views of where the product stood. The new CEO wanted to know who was right.
I told him I wouldn’t believe either of us if I were sitting where he was sitting. We were both too dug in. What I suggested instead was that he go out to the rest of the organization… the sales engineers, the consultants, the support and development teams… and ask them. They were the ones closest to the customers. They knew where the product actually stood. Whatever he found out there would tell him what he needed to know, and then the decision would be his.
My family thought I’d lost my mind when I told them the story. You could have gotten yourself fired, they said. Why didn’t you just smooth it over?
Because that wasn’t the truth. I had built relationships with customers in the field and I felt a responsibility to them, to myself, and to the company to say what I actually believed no matter how it ended for me. And to his credit, the CEO didn’t take the easy route either. He went out and did exactly what I suggested. He called together the people closest to the work and he listened to what they told him.
In that moment I recognized something. This wasn’t agreement. He might ultimately decide I was wrong. What mattered was how he went looking for the answer.
Not through hierarchy or politics but through the people closest to the work.
That’s when I realized we were coming from the same place. Not the same title. Not the same position. The same place.
I’ve been thinking about that conference room a lot lately. Partly because of a post from Geoff Woods that’s been making the rounds… the argument that a CEO can’t cast a vision for something they’ve never touched. He’s right. But I think he stops one rung short.
I keep going back to Jerome Kanter. In 1967 he wrote a book called The Computer and the Executive. It came from my late father’s library. One of the questions on the back asks what the human relations impact is of bringing in a computer. Substitute AI for computer and the question is identical to the one every organization should be asking right now but mostly isn’t.
Pentland and Agrawal put it plainly in the Digitalist Papers (Susskind, Stanford Digital Economy Lab, 2025)… enormous investment in developing AI, comparatively little invested in understanding what it does to the systems it enters.
That’s not a technology problem. It’s a leadership problem that starts at the top and compounds all the way down.
The same failure repeats across every corporate function. The tool arrives, the governance doesn’t, and the human judgment that made the function worth having walks out with the last round of layoffs. The implementation differs. The pattern doesn’t. Nowhere is it more visible right now than in hiring.
A 2026 HBR study analyzed over six thousand first round screening sessions and found that the early hiring funnel is breaking on both ends simultaneously. AI optimized resumes have lost their signal value at the top. Live video interviews are being gamed by real time assist tools at the bottom. What’s left in the funnel in the middle isn’t a talent signal. It’s a desperation signal. The people willing to engage with an AI screening bot aren’t necessarily the best candidates. They’re the ones who felt they had no other choice. The talent that walked away shaking their heads… that’s the loss nobody is measuring.
And it gets worse. A 2025 University of Maryland study found that AI screeners show self-preference bias… consistently favoring resumes that resemble their own outputs. Candidates whose resumes matched the evaluating model’s style were up to sixty percent more likely to be shortlisted than equally qualified candidates who wrote their own. The filter isn’t finding talent. It’s finding mirror images of itself.
There’s a voice I keep coming back to on this. Bryan Creely runs a YouTube channel called Life After Layoff. He spent twenty years in corporate recruiting. He makes the point that a good recruiter advocates for you. Goes to the hiring manager and makes the case for the candidate who doesn’t check every box on paper but has something worth betting on. Builds credibility over years and uses it to open doors.
He also makes the point that when you see recruiters getting laid off, that’s your canary in the coal mine. The people whose singular job was to find talent are gone. The function remains. The judgment doesn’t. The algorithm inherited the role but not the wisdom that made the role worth having.
That’s not a hiring problem. That’s a Kanter problem. Nobody asked what the human relations impact would be.
That’s why I’ve become increasingly skeptical of a certain version of AI fluency. The kind that mistakes touching the tool for understanding the terrain.
Sherpas are not tour guides. They are the load bearing human infrastructure of the entire expedition. Oxygen management, route finding, weather reading, rescue capability, institutional knowledge of the mountain built over generations. Without them the climbers don’t summit. Without them the climbers die.
The puja before a Himalayan expedition isn’t a photo opportunity. It’s the moment the expedition acknowledges what the mountain actually requires. The respect. The preparation. The humility to understand that the mountain doesn’t care about your credentials.
A CEO who picks up an AI tool for an afternoon and calls it fluency has performed the ceremony without internalizing what it means. The puja was ceremonial. The mountain is still the mountain.
Fluency matters. But the CEO who goes out to the people closest to the work… the ones with customer contact and field scars and the kind of reputation that doesn’t show up on an org chart… and actually listens to what they find… that CEO is asking Kanter’s question. That CEO is doing the puja right.
The CEO didn’t trust the title. He didn’t trust the org chart. He didn’t even trust the loudest voices in the room. He went to the people closest to the work and listened.
Nearly sixty years after Kanter asked what happens when new technology enters a human system, the question hasn’t changed.
The leaders who navigate AI successfully won’t be the ones who perform fluency. They’ll be the ones who know where to look for the truth.