How do executives use AI to move faster toward the CEO seat? They use it to multiply output before they hold the title that normally justifies a bigger team. Gregory Shepard, CEO of Startup Science, has watched AI compress a six-person marketing function down to one person and a ten-person engineering team down to the same. That kind of output puts you in strategic conversations years before the org chart would normally allow it.
Quick Answer
Executives get the most value from AI when they configure it to challenge their thinking, not confirm it. Glenn Gow, The Scaling Executive Coach, works with CEOs who instruct their AI models to argue against a decision before they finalize it. The strongest executives delegate analysis and pattern recognition to AI while keeping judgment for themselves. This habit catches blind spots before a board meeting does, not after one.
The Sycophancy Problem Every Senior Executive Faces
Most AI for executives agrees with whoever is typing. David Simnick, CEO of Soapbox, ran into this early and built a workaround into how he uses the tool for big strategic thinking. He told me the goal is a model tuned to be objective, “not having something that just repeats back what it thinks you want to hear.” He toggles his settings down until the answers get uncomfortable on purpose.
Tom Alexander, CEO of Holistic, takes the same problem and applies it to leadership instead of strategy. He builds AI tools that survey a company’s culture and hand a CEO results without softening them, including on the question of whether the CEO is actually living the values he claims to hold. Tom put it this way: “I can’t ask myself this because I’m going to fudge my own answer, but I can ask a computer that, it will give me a dispassionate answer and then I’m uniquely suited to use that information for progress.”
Glenn Gow tells the CEOs he coaches to schedule that toggle deliberately, on a calendar, before every quarterly planning cycle. Waiting until the mood strikes turns a discipline into an accident. The executives who treat objectivity settings as a recurring appointment get the uncomfortable answer on a schedule they control, not the week a board member forces it on them.
Glenn Gow describes this as the gap between ai for senior executives and AI for everyone else in the company. A junior employee uses the tool to move faster on a task someone else assigned. A CEO uses it to find out where his own judgment is wrong before a room full of people watches him find out the hard way.
Building an AI That Argues Back
Philippe Bouissou, CEO of Blue Dots Partners, LLC, builds AI specifically to provoke CEOs rather than serve them. He described the product to me directly: “This is not an agent. This is not a bot. It doesn’t really do anything, but it provokes the thinking of the CEO within the secured context of the company.” Glenn Gow has watched CEOs delay building anything like this for a full year while they wait for a vendor to ship one instead.
Philippe gave me two examples of what that looks like across a single week. The first: “I have a board meeting next Friday. Here is my deck. What are the three most difficult questions the board’s going to ask me and how should I answer?” The second sits even closer to the decisions that define a CEO’s tenure: “I have three companies I have in mind to acquire, which one should I acquire and more importantly, why. Or I’m recruiting a CFO, I got 75 resumes, I can only meet with five. Who should I meet with and why?”
Glenn Gow sees most executive education on AI treat the technology as a productivity add-on, something you bolt onto existing habits. The CEOs who gain a competitive advantage treat it as a standing member of the room where the hardest calls get made, the way Philippe does. That distinction shows up before the acquisition closes or the CFO gets hired, not after.
Philippe’s approach only works because he anchors every question to a live decision instead of a vague prompt. Glenn Gow sees the same difference in his own coaching calls. A vague question about growth gets a vague answer. A question built around one specific number gets an answer a CEO can act on that same day.
What You Hand to AI and What You Never Hand Over
Every CEO in this brief draws the same line, even when they describe it in different words. Marianne Abib-Pech, Managing Partner at Transitions First, states the boundary directly: “I think me personally, will delegate analysis, pattern recognition, challenge to AI, but I will never delegate the judgment part.” Glenn Gow enforces the same boundary with every CEO he coaches, no matter how sophisticated the AI model gets.
Gregory Shepard frames the same split around a supply chain instead of a decision. He described data as raw material that becomes information once it gets structured, then named wisdom as the final, distinct stage: “That wisdom component you need people for still. But the data component and the structuring component you can do with AI.” Below is how that division plays out across a typical strategic decision.
| Function | Who Handles It | What the CEO Gets |
| Scenario modeling and headcount math | AI | Fast answers without exhausting the ops team |
| Cultural and organizational survey analysis | AI | A dispassionate read no employee will give a CEO directly |
| Pattern recognition across deals or resumes | AI | A shortlist built on structure instead of gut feel |
| Final judgment on the acquisition, hire, or strategy | The CEO | Accountability for the call, informed by the data above |
That table only works because Marianne and Gregory hold the line on the bottom row. An executive who lets AI make the final call has stopped leading and started rubber-stamping.
Rehearsing the Hardest Conversations Before They Happen
Steve Harmon, CEO of Spartan Logistics, turned ai upskilling into a leadership directive instead of a course. He gave his team an unusual mandate: “I want all of you to not Google anything for the next quarter. I want you to only use the chat bot of your choice.” He followed it with a second instruction that turned the tool into a coaching asset: “I want you to build out a personal advice board for your role. I want you to use it for reviews with clients and with your team members.”
David Simnick uses the same rehearsal instinct on a specific, high-stakes retail conversation. He described asking his model: “How do I potentially position this conversation with the vice president of merchandising so that we can get these end caps?” He runs that scenario before the actual meeting, not during it.
Marianne Abib-Pech takes rehearsal a step further. She built what she calls a synthetic think tank, an AI model she uses to stress-test a new idea before she takes it to real people: “I actually created a synthetic think tank, which was AI based, where I could start my thinking, stress test my idea.” Only after that internal argument does she bring the idea to her real, human advisors.
Search for the best ai courses for executives and you will find hours of generic video training built for a general audience. Steve Harmon did not send his team through one. He sent them straight into client meetings with the tool already open.
That order matters more than it sounds like it should. A course teaches you what the tool can do in the abstract. Steve Harmon’s quarter-long ban on Google forced his team to find out what the tool could do with a client on the phone, not a demo on a screen. The lesson stuck because the stakes were high the first time they used it, not the tenth.
Using AI to Correct Your Own Blind Spots
Kriti Sharma, CEO of IFS Nexus Black, uses AI for the work you would expect from a former software engineer. She runs board papers, customer outreach, and strategy documents through coding agents. But she told me the surprising part is different: “A more of a hidden superpower that I recently uncovered is the ability for AI tools to help me with emotional intelligence… to be able to see the other person’s perspective.” Glenn Gow has started asking his own coaching clients where in their week an AI tool could give them that same kind of read on a colleague, not just on a spreadsheet.
Tom Alexander’s dispassionate survey results only matter if the CEO reading them is willing to act. He named the trap directly: people, including CEOs, are “horribly unreliable narrators about themselves.” Glenn Gow builds this same check into his own coaching process before a client ever runs a formal survey.
Glenn Gow watches this play out in coaching sessions more than any other single pattern. A CEO will read a dispassionate survey result, agree with it out loud, and then quietly explain why his situation is the exception. The tool did its job. The willingness to act on an uncomfortable answer is still a human decision, and it is the one most executives skip.
From Data to Wisdom: Where the Machine Stops
Chris Crowe, CEO of CMBYND, warns that most people treat AI output as settled fact instead of a starting point. He put it bluntly: “The vast majority of people take what is produced from AI as gospel. They don’t question it.” His fix was structural. He rebuilt his own firm around the problem, moving away from the traditional pyramid where junior staff sit at the bottom doing analysis: “When AI came around, I realized that we could actually build a diamond instead of a pyramid.”
| Model | Definition | Ceiling |
| Pyramid | Knowledge workers at the top, a large supporting team doing analysis at the bottom | Growth is capped by how many analysts you can hire |
| Diamond | AI absorbs the supporting analysis layer, knowledge workers sit closer to the center | Growth is capped by judgment, not headcount |
Chris told his own team the shift required was personal before it was organizational: “You need to turn shift from a doing to a thinking posture.” Glenn Gow asks every CEO he coaches to name one AI output from the past month they accepted without pushback, then explain why.
None of this argues against speed. The force-multiplier effect Gregory Shepard described earlier in this piece is still true and still available to any CEO willing to use the tool that way. The argument is about sequence. Speed without a model built to challenge you produces fast, confident, wrong decisions at scale. Speed built on top of an objective, argumentative AI produces fast decisions a board can actually defend.
Glenn Gow sees this pattern across nearly every coaching conversation he runs with CEOs adopting AI. The executives who advance fastest let the tool disagree with them first, then make the call themselves.
FAQ
How should a CEO start using AI for strategic decisions?
Hand over analysis first and keep judgment for yourself. Marianne Abib-Pech, Managing Partner at Transitions First, delegates analysis, pattern recognition, and challenge to AI, but after she stress-tests an idea inside her own synthetic think tank, she still brings it to a real think tank of people scattered around the world before deciding anything. Glenn Gow calls this the correct order: analysis first, argument second, decision last.
How do executives stop AI from just telling them what they want to hear?
Toggle the model toward objectivity before asking it anything important. David Simnick runs the same strategic question through several different AI models before trusting any single answer, specifically so the tool cannot just repeat back what he wants to hear. Glenn Gow has watched CEOs skip this step and pay for it in the boardroom.
Should a CEO ever let AI make the final call?
No. Tom Alexander argues AI offers a capacity to be the best version of yourself that you could never accomplish on your own, but that capacity only matters if a human still makes the final call. Marianne Abib-Pech draws the same line from the investor side: analysis and pattern recognition go to the machine, judgment never does.
What is an AI thought partner and how does it differ from a chatbot?
A thought partner questions you instead of completing tasks for you. Philippe Bouissou, CEO of Blue Dots Partners, built exactly this distinction into his own AI product: it does not act as an agent or a bot, it exists only to provoke the CEO’s thinking inside the company’s own context. Glenn Gow uses the same standard when he evaluates any AI tool a client wants to add: does it answer, or does it push back?
How can AI help a CEO model hiring and headcount decisions without burning out the team?
Gregory Shepard, CEO of Startup Science, runs his own scenario modeling instead of routing it through operations. He described testing a specific headcount and salary trade-off out loud: “What happens if I hire five people at this much?” He added that this kind of repeated modeling could exhaust someone in a shared services role, but a CEO can run it as many times as he wants and get good results without pulling staff off other work.
What does effective ai upskilling look like for senior executives?
It looks like real work, not a training module. Kriti Sharma, CEO of IFS Nexus Black, described her own routine this way: “I work like a software engineer now. I have my daily knowledge work, like whether it’s board papers, it’s customer outreach, it’s emails, it’s pipeline reviews, strategy documents. I put them in GitHub. I work with coding agents to be able to push the boundaries live in meetings.” Steve Harmon reached the same place by banning his team from Google for a quarter.
How does AI let a CEO take on more complex problems without adding headcount?
Kriti Sharma frames it as a scaling question, not a headcount question: “I believe now with the superpowers that we have using AI to scale organizations, to scale capabilities, to scale skills, we can solve multi-dimensional problems as long as we stay focused on the same few customers.” Glenn Gow tells clients this is the test to run before hiring: can the current team absorb one more layer of complexity with AI, or does the complexity require a new hire regardless.
How is AI changing the structure of consulting and professional service firms?
Chris Crowe, CEO of CMBYND, rebuilt his firm around a different shape entirely. He explained the old model first: “Traditional organizations in consulting were built like a pyramid where you had your knowledge workers at the top and at the bottom you had your supporting team that’s helping you do analysis.” Once AI could handle that supporting layer, Chris moved his firm to a diamond shape instead.
How should a CEO push a leadership team to use AI beyond routine tasks?
Set a standing, personal use case, not a one-time training day. Steve Harmon told his team, pointing at the AI tool he had just put in their hands: “This thing is capable of anything. I want you to build long-term projections. I want you to think strategically.” Glenn Gow tells clients the instruction only works if the CEO models it first, on his own calendar, before asking a team to follow.
What is the biggest mistake executives make when they first start using ai for executives tools?
They turn the objectivity back off. Glenn Gow watches CEOs discover a model that argues with them, get uncomfortable with the friction, and slowly nudge the settings back toward agreement over a few weeks. The value of any ai for executives tool disappears the moment it starts being nice again.
Does using AI actually help a CEO reach the seat faster?
Yes, but only if it is used for judgment, not just speed. Gregory Shepard’s force-multiplier numbers from the top of this piece are real, and Glenn Gow sees CEOs move into bigger roles faster when they use AI that way. But Glenn Gow is direct about the trap: speed without a model built to challenge you just produces bad decisions faster. A board promotes the executive it trusts with the call, not just the one who moved first.
Is it safe for a CEO to put confidential company data into an AI tool?
There is real risk here, and it depends on the tool and the data. Kriti Sharma, CEO of IFS Nexus Black, routes board papers, customer outreach, and strategy documents into coding agents as part of her daily routine, which is the kind of decision that has to be made deliberately, tool by tool, not by default. Glenn Gow tells clients this is a conversation to have with their own IT or security lead before rolling AI out further. No CEO should treat it as a blanket yes.
What is the biggest risk of relying on AI too much as a CEO?
The biggest risk is accepting the output before questioning it. Chris Crowe has pointed to roughly 75 percent of people accepting AI output as fact without checking it, the same instinct that lets a bad recommendation reach a board unchallenged. Glenn Gow treats that number as a warning specifically for CEOs, since a wrong call at that level costs more than a wrong call anywhere else in the company.
What You Do Next
Every executive in this piece got further, faster, by handing analysis to AI and keeping judgment for themselves. If you want help figuring out where that line sits in your own company, book time with Glenn Gow directly and bring the decision you are least confident about right now.
