Editor’s Note: Keith Ferrazzi is the opening keynote at our Leadership Conference in San Antonio, November 4–6. The whole program is built around speed—the tools, alignment and accountability for more disciplined execution in complicated companies. Former Sen. Phil Gramm will be there to unpack the midterms and help you navigate the economy, and Bob Nardelli, who ran Home Depot and Chrysler, has the closing keynote. Join 500+ CEOs at the Hyatt Regency Hill Country—and go faster in 2027. Learn More and Register >

From his offices in the Bay Area, Keith Ferrazzi has an unusual vantage point on leadership in America’s businesses. He spends his weeks inside the executive teams of some of the largest companies in the world, and he spends his evenings with the founders trying to eat them. “I’m on the two ends of the dumbbell,” he says.

What he’s hearing about from the big end is frustration—especially around AI, which dominates the conversation. Boards want returns, CEOs have spent two years splurging on tokens and the numbers aren’t showing up.

Ferrazzi’s diagnosis is that, when it comes to AI, most companies are measuring the wrong thing entirely, and the thing they should be measuring can’t be done inside the org chart they already have. Which is where his recent book, Never Lead Alone: 10 Shifts from Leadership to Teamship, comes in.

We talked in early September. The conversation has been edited for length and clarity.

What’s the vibe out there right now with the CEOs you talk to and work with? What’s keeping them up at night?

Getting real returns from AI is still more elusive than they’re comfortable with. They’re getting a lot of pressure from boards to be cracking the code and at the same time swinging for the fences. There’s an expectation that by now we would have gotten to figure it out and gotten some real returns.

And then the second piece is how to get their teams to get there. Because the reality is the old ways of working on this kind of transformation are not working.

Where does that start? What’s the big thing you see CEOs getting wrong?

Let me describe how a CEO should talk about AI. And right now we’re muddling it together. Number one, a lot of big companies, they’re defining AI by how many people are using the tools. Bad idea. It’s not about how many people are using tools. All that will do is make the LLMs happy by running up a big token usage.

There are three areas, and CEOs are meshing their conversation around AI and they’re not breaking it into those three chunks. One area is productivity improvements. The next area is zero-based workflow redesign. That’s the big unlock for ROI. The final area is using AI to re-engineer your product itself, and a lot of companies have been doing that for a long time, because AI at the CTO-product level has been around longer than generative AI.

The middle one is where you’re going to win the most.

Start with the first one. If tool adoption is the wrong metric, what’s the right one?

You should not invite employees to just get on their journey to use AI. You should get them excited to be on a journey to get results from AI. That’s where you should be focused.

Anybody who raises their hand, call them a yellow belt. We love the idea that somebody’s interested, you’re a yellow belt. But for you to get celebrated, I want you to become a brown belt. I want you to show us in demonstrated small five-minute videos or town halls where you raise your hand and you say, here’s what I’ve been doing. Five minutes. Here’s the application. I’ve used AI. Here’s the results I’ve got. Don’t let somebody talk just about the application unless they talk about the results they got. Here’s the application. Here’s the results that I got. Here’s three things that I learned, and one thing that I struggled with and overcame. That’s a five-minute video. That’s a five-minute stand-up. Those are brown belts.

The black belts are those who have been able to take the AI and not only make what they’re doing more productive with ROI, but also start re-engineering their own individual workflows down at the individual level. Black belts are the ones who can turn around and teach the brown belts how to be black belts. The brown belts should teach the yellow belts how to be brown belts. People in the bowels of your own organization should be teachers.

So that’s a system around productivity that is better than token usage.

The theme we keep hearing from CEOs is that they’ve got a strategy, they just can’t get everybody to move as fast as the world is moving. And that’s what your book is about. How does Teamship come into play here?

The next layer of AI usage is re-engineering functional and divisional workflows. Really not at the level of a person doing the workflow redesign, but it’s like, okay, how are we re-engineering supply chain? How are we re-engineering go to market? And that is where you need a competency as a CEO.

The big problem you have is that you do not have that competency in-house. More importantly, if you did, you’re too bound to org charts. The reality of re-engineering workflows is across the org charts. That’s where that book comes in.

Chapter number one is teaching you the team has nothing to do with your org chart. Companies have to realize they need to start organizing around transformation, not organizing around org charts and functions. What are the four big rocks we’re trying to move? And within each one, what are the critical initiatives? That’s what you organize around, and those cut across the silos of an organization. The book is your roadmap on how to work more effectively in these cross-fertilized teams.

Once you’ve got that team pulled together across functions, what has to be true about how they work?

You need to have a social contract across these matrixed teams in your organization that agree to a new way of collaborating, that is egoless, that is bold and trying to hit the ball, hit it out of the ballpark, not just do incremental change, that is willing to be candid and create a challenge culture in the dialogue.

So if you’re re-engineering the end-to-end supply chain and you’ve got people from procurement, people from sales, people from manufacturing, people from technology in this group—the bullshit that we have to get rid of is “Who’s the leader?” The reality is this is a group of individuals that need to learn how to have a social contract with each other, that can challenge each other, that can point out where there’s a need for each other to level up, that can be willing to be egoless and take my entire piece of this thing out of the process and identify it.

That is a different social contract than the bullshit of conflict avoidance, the bullshit of being in silos, the bullshit of maximizing for my piece, and the bullshit of incremental change instead of true transformation—because they don’t even know what AI could do, because they’re not even imagining how amazingly different the world could be, because they’re in their own vantage point of what it always has been.

What are you telling CEOs about how to get started?

What I’m teaching them is get it right somewhere before you try to get it right everywhere.

Last night I had a group of CEOs in the Bay Area, and people were bemoaning the lack of advancement that they were getting. And I was talking to one guy, and I said, “Listen, where do you need the most lift?” He’s like, “Go to market.” I said, “Great. You now run go-to-market transformation.”

In other words, make sure that you put your superpower of being a CEO behind what is most important. And I want you to work on monthly agile sprints where they come to you and the entire executive team and we re-engineer the go to market.

Number one, you need to say that I want not 50 percent more sales productivity. At a minimum, it’s got to be 2 to 3x, maybe 5x sales productivity. How do you define a new go-to-market process that’s going to give you 2, 3, 5x sales productivity?

Once you put that audacious goal out there and the parameters around whoever is needed to be in that room are on the team—that could be product, it could be marketing, it could be sales, it could be technology, it could be whatever it needs to be—that’s the group that has to fix this go-to-market problem, not the sales organization. That is the team that’s going to do the 2x, 3x revenue acceleration.

A lot of our readers run mid-market companies. They don’t have the big staff. They see the opportunity, they know where they need to go and they can’t seem to get there. What’s the takeaway for them?

You need to own the landing of successful AI re-engineering in a particular area. Every month the team is going to come together in front of you, with you, and you are going to be the driver of not letting anything stand in the way.

On a monthly basis, the head of supply chain, the head of sales, the head of whatever is slowing you up. We are stress testing—chapter three of the book. We’re stress testing the results we’ve gotten, where we’re struggling, where we’re going and nobody is off the hook. We are all collectively responsible.

So you’ve got to be 10x bolder. Zero-based workflow redesign. You’ve got to be the driver of the critical change. If you just step back and say, “I’m asking all of you to raise the water levels,” you’re going to get nowhere.

What are you learning from the small, AI-native companies that you’re bringing back to the big ones?

What I’m learning from the kids is their nimbleness, their agility, is defined by the process of how they collaborate. They don’t use meetings as their primary form of collaboration. They use asynchronous collaboration. Everybody’s point of view is known before you show up in a meeting, so that you show up in a meeting and you can wrestle the clear disagreements and land the plane in one meeting.

The average big company starts their collaboration in the room. Half the people are heard, half the people are not. Then there’s back channel, behind the scenes conversations. There’s multiple meetings to get people’s voices heard. And it’s just 10 times the cycle time of these kids.

You wrote recently wrote about bringing agents into the workflow. Where are you seeing this actually work?

The ones who are making it work are the ones who are hiring these kids, who have picked a solution set—like call centers—and have said, we’re gonna fundamentally re-engineer call centers. These kids out of Y Combinator, these AI native kids who are like, “Let me look at a call center and let’s figure out how we build AI to get a higher net promoter score with 20 percent of the people.”

These kids are not the big firms. These are AI-native thinkers. Not the ones who are our big old school consultancies and systems integrators who are now trying to rebrand themselves.

The biggest breakthroughs that these CEOs last night were reporting were from AI native interns that they were bringing into the company. It was shocking. I had never heard that before. Like, they were bringing in these AI native, swashbuckling kids into the company, and creating these digital services groups that were swarming in, working with the business unit leaders. And really bringing an entirely different way of thinking.

Last thing. Where does the belt system fit against all of this—is it core, or is it running alongside?

That’s a nice to have. That is a nice to have thing that your head of HR, your head of training and your CIO should own and run with that initiative, which is the elevation of everyone’s productivity as much as possible.

I’ll be honest, here’s what you do. You invite the world in to do that. You see who gets to be a yellow belt. You see who gets to be brown belt. You see who gets to be black belt. You make sure that there’s plenty of time for everybody in the organization to see real benefits from AI from their peers and be invited to go on the journey.

I’ll be crass for a second. Those who after six months, maybe after a year, those who have not gotten on that journey—fire them. They’re yesterday’s employees.