So this is my take. AI is just not going to replace management and mentoring. What's really happening here is that a group of CEOs who overhired want to juice their valuations and engage in AI washing. Management is an ability to soften the chaos created by leaders. The notion that your AI is going to bring everybody in and they're not going to need to have contact with anybody else, and the AI will manage them, tell them what to do, and at the end of the year give them their performance review and a bonus — even if that was possible, which I don't think it is — it's a nihilistic, [__] up vision of the world.
In today's office hours, we discuss AI replacing managers, the rising cost of enterprise AI, and how to become a better storyteller. Question number one: first question comes from a Reddit user, MX AGNC asks, "We all know junior and middle management roles are being decimated by AI, but what might be next? Experienced independent contractors, senior leadership, or even CEOs? There seems to be different cases for each and would love your thoughts."
Yeah, that seems to be a big topic lately that AI can replace all management layers. Gartner projects one in five companies will eliminate more than half their middle managers by end of this year. By the way, Gartner doesn't know what the [__] they're talking about. I just don't — I don't buy that. Anyways, Coinbase cut 14% targeting pure managers. Microsoft laid off 6,800 citing AI and flatter org goals. Here's what Jack Dorsey had to say about his company Block and layoffs. Open quote: "The first time a system can maintain a continuously updated model of an entire business in ways that previously required humans relaying information through layers of management."
So this is my take. AI is just not going to replace management and mentoring. What's really happening here is that a group of CEOs who overhired want to juice their valuations and engage in AI washing. Management is an ability to soften the chaos created by leaders. They're the ones that actually remember why this actual [__] company works. And as you grow, you need people that these people can roll up to in terms of mentorship, job descriptions, solving problems. And what happens is eventually over time, people get hired sometimes out of their competence into management roles. And you end up with fatty deposits everywhere. And it's a natural, healthy part of the cycle to go through and cull the ranks not only of managers, but of some doers, if you need less done in that field or they're not that productive.
The notion that your AI is going to bring everybody in and they're not going to need to have contact with anybody else, and the AI will manage them, tell them what to do, and at the end of the year give them their performance review and a bonus — even if that was possible, which I don't think it is — it's a nihilistic, [__] up vision of the world. And that is okay. Even if you could figure out a way for one person to own everything and create so much shareholder value that you could, I guess theoretically, give people enough cheap calories and entertainment — bread and circuses — so they could just watch Netflix all day and be fairly happy, which they wouldn't, because work gives people purpose.
I don't think that's going to happen because I work in a company that's growing, that's considered a new economy company. We use AI on almost everything. But the last mile involves people and it involves judgment. And the idea of me sitting down and outlining job objectives, providing feedback at the end of your review, bonus, solving conflict with AI — I just don't — I don't think anyone's going to go for it. And also, AI gets it wrong all the time.
And purely AI-driven businesses. AI is essentially predicting the seventh word after looking at the first six and coming up with patterns such that it can answer any question or find any data and spit back a reasonable facsimile of a correct answer to any question. But it's regression to the mean. And that is, it gives you great data on a question. But if you're asking it to brainstorm, if you're asking it to solve a problem, if you're asking it to manage somebody who clearly has a substance abuse issue, it can give you facts and best practices, but everything's driven to the mean.
The most successful companies will be companies that I believe that AI would say this will never work because they're thinking unconventionally. So I don't — I just don't buy. We always have these fever dreams of flattening the organization and reducing costs. My father was laid off in the 80s in this big kind of wave in corporate America called re-engineering, where people basically — after World War II, we had a monopoly on all manufacturing globally. And as a result, companies just kept growing and hiring more people for no real reason. And they woke up in the 80s when Japan was starting to come online and the [__] were fat and happy. And they essentially fired anyone that had the title VP in it. And they found they didn't miss them much.
Anyways, yeah, it's healthy on a regular basis to go through the company and cull and figure out where you can delayer, have fewer reporting layers. I get it. This again is individuals who quite frankly, in my opinion, don't know a lot about management and think that technology or the capital can replace people. And there will be, I think, job destruction, management destruction across certain industries. I don't think it's going to happen as fast as anyone thinks. And I hope and I believe that the opportunities AI will provide in terms of an increase in productivity, margin, profitability, shareholder value, and the wealth and capital to go start new businesses that leverage AI — that hopefully will create as many jobs if not more than are destroyed.
The key question is: will the velocity of decline overwhelm our ability to repair and renew and create new jobs? That's the question. And unemployment actually isn't the issue in terms of social unrest. What causes social unrest usually is not someone who can't find a job, but someone who's working 16 [__] hours a day and still can't afford to buy, you know, meat. Those are the people that tend to get really, really angry. So I don't — I think this flattening of the organization is a bunch of people who overhired during the pandemic and also maybe overestimated the demand for their product or service.
Want to sound revolutionary. I'm a big fan of player-coaches, and every company I've had, everybody has to do something. I do a podcast in addition to managing. The woman who runs our company, Katherine Dylan, manages a lot, but she also edits things, is fantastic in terms of design. She gets very involved inside. She's — everyone does, if you will. I think that's a good generalist within a company. But you absolutely have to have managers. You have to have people communicating who create institutional and corporate knowledge.
So in sum, I think this is more nihilistic [__] from a group of people who have no or very little humanity, are very nihilistic, and also in my view don't understand or aren't willing to admit that, you know, managers evolved for a reason. And a good manager motivates people, figures out where the overlap creates a whole that's greater than the sum of its parts. So will there be a culling? Yes. Will there be a lot of quote-unquote middle managers? 100%. And they're the ones, quite frankly, that understand why we're all [__] here.
Support for the show comes from NetSuite. Every business is asking the same question: how do we make AI work for us? The possibilities are overwhelming, but sitting on the sidelines is not an option, since you know your competitors are already making their move. No more waiting. With NetSuite by Oracle, you can put AI to work today. NetSuite is a top-rated AI cloud ERP trusted by over 43,000 businesses. It's a unified suite that brings your financials, inventory, commerce, HR, and CRM into a single source of truth.
Now, with NetSuite AI Connector, you can use the AI of your choice to connect to your actual business data and ask every question you ever had — from key customers to cash on hand to inventory trends. Plus, you can automate those tiresome manual processes. Let's see your competitor do that. This isn't another bolted-on tool. It's AI built into the system that runs your business. Whether your company earns millions or even hundreds of millions, NetSuite helps you stay ahead of the pack. If your revenues are at least in the seven figures, get our free business guide, Demystifying AI, at netsuite.com/profg. The guide is free to you at netsuite.com/profg. netsuite.com/profg.
Question number two comes from Aaron who emailed us: "Scott, I'm a strategy lead at a Fortune 10 retailer and my AWS agent usage on Claude is probably running a few hundred in tokens a day. Multiply that across an engineering org and you've got seven or eight figures a year fast. Companies are simultaneously telling employees to use AI for everything and quietly watching the meter spin. How does this resolve? Do enterprise AI bills become a real P&L line that competes with headcount?"
That's a good question. So most individuals pay a flat monthly fee for AI tools. But when companies build AI into their own workflows — things like coding assistance, automated pipelines, internal agents — they pay per use with no ceiling. And the more engineers that use it, the bigger the bill automatically. In other words, modern enterprise AI is metered, similar to electricity or mobile data. The number of organizations that reported AI costs as a serious budget concern doubled in a single year, from 31 to 63%.
Now the question is: will AI just get cheaper? Partly. The cost of a basic AI query has dropped roughly 50x in three years, but the more powerful the task — including complex reasoning and AI agents running autonomously — the more compute it burns, and those prices haven't followed the same curve. Think of it as Moore's law, where processing power doubled every 18 months and the cost got halved, but people continue to find more and more uses for bandwidth or for processing power. And the same thing's happening here: that these technologies are usually highly elastic, and that is, as the price of compute plummets, you'll see more uses for AI and it'll broaden the market, and there'll still be more demand for compute and for inference.
Is this becoming a real P&L line? Oh, 100%. 59% of companies are investing over $1 million annually in AI, but only 29% report significant ROI. So right now the investment is ahead of the return, and in practice the usage costs are frequently not the dominant expense once you add integration, security, and change management. The full cost of AI is much larger than what shows up on the bill.
Welcome back. Question number three: "Hey Scott, my name is Andre and I'm from Edmonton, Alberta, Canada. My question to you is related to a point that you've made in several podcasts, and it's the importance of being able to craft a strong narrative and being a good storyteller. My question to you is: what tips do you maybe have for either content creators or business professionals or even professors to be able to craft a good narrative? I notice in a lot of your books and in a lot of your podcasts, you're really big on research and leading with numbers and facts. What other tips do you have for individuals who really want to strengthen this skill?"
So I would refer you at least first off — Prof G Markets' research lead did a live stream on Substack on this very topic. She puts together my decks for our speaking gigs, and she put together a deck on storytelling, and it's still, I think, to date the most popular live stream we've done on Substack. And this is kind of topline, the cliff notes on what she said. You want to create drama. Drama is what makes people care. And there's three ways to do it. You make it emotional. Even unsexy topics can evoke anger, sadness, or excitement if you find the human stakes.
I talk a lot about my personal life and things that have moved me emotionally, whether it's my kids. You know, I talk a lot about my parents, and, you know, I'm not afraid to sort of be quite vulnerable and talk about, you know, success. And I mean, I try to be pretty raw up there. And I find that of all the things I've ever done, the most popular newsletter I've ever done was when I talked about putting down my dog and just the emotions I was feeling. I write about stocks and I'm a quote-unquote business analyst, or write books about economics, and the most popular piece of content I ever put out there was about my dog.
Most recently, the clip I had go viral that got, I don't know, 400 or 500,000 likes and I think 7 or 8 million views was on Diary of a CEO. I was talking about how much I miss my mom. So emotion. Right. Surprise: unexpected information triggers dopamine, which literally enhances memory formation. And then kind of wow factor: data. Before adding any data point, ask: would this make someone say, "I didn't know that"? If not, cut it.
You also want to zoom out. Don't settle for the obvious first-order take. Ask what the upstream causes and downstream effects are. Let's look at GLP-1, for example. Instead of covering Novo Nordisk, we followed the supply chain to Catalent, then followed patient behavior all the way to a spike in plastic surgery. That's where, you know, that's where the fun starts: is, okay, we know that GLP-1 drugs are big, but what else is going to be big?
Try to provide context. Big numbers mean nothing in isolation. Data center water use sounds alarming until you learn one data center uses half as much water as the average golf course per year. So context is what makes data meaningful rather than fear-mongering. And then aesthetics, visualizations. People process images faster than words. We look through every deck and say, "How do we show this versus say it? How do we replace it with visuals?" The Nvidia versus Intel market cap example — you can say it in a sentence, but showing it lands harder.
We do these treemap graphs that show spatially market cap of, you know, Nvidia is greater than the market cap of Germany. And you show every single public company, and that one square is bigger than the other. So some other just tactical things. It sounds very basic. I think great storytellers are usually pretty big readers at a young age. I didn't read an inordinate amount, but I did read quite a bit when I was younger.
Also, if you want to be a great storyteller, I think you have to be a competent writer. I think the discipline of trying to write out grammatically correct prose, organize it, chunk it. You know, Jeff Bezos is famous for anyone who pitches an idea has to do it in a five- or 10-page deal memo. I think that — and by the way, I got a C in English my senior year in high school. I found writing very difficult. I now technically communicate or write for a living. But I think if you want to be a really great communicator, you have to be at least a competent writer.
Then what I want you to do is I want you to pick a medium that you feel comfortable in. It might be LinkedIn. It might be a podcast. It might be videos. It might be presentations using PowerPoint. But I need you to find a medium or a platform that you think you could be in the top 1% of and try and get to the top 1%. And that is, say, all right, I'm going to really be an outstanding orator, or I'm going to commit to writing amazing content on LinkedIn, and I want to be in the top 1% of LinkedIn content creators.
What does success look like? Success is small acts of discipline every day that add up and compound, similar to investments. And to be a great storyteller, I want you to read a lot, I want you to write a lot, and then I want you to find different mediums where you think you could be in the top 1% and start investing in those mediums. And have sort of a ready, fire, aim: I'm going to do three LinkedIn posts a week. I'm going to try and amass 10,000 followers on Instagram by a certain time, right? Whatever the medium is.
Or my goal is to get — I want to start a Substack and I'm going to put out at least one 1,500-word article on this topic every week. So find your medium, set deadlines, set goals and benchmarks, and just start. But if you're not a great presenter — everybody thinks a great communicator is someone who gives this moving eulogy. Well, you might be the person who's really good at handwritten notes. You might be the person who's really great at doing 30-second TikToks. Find your medium. Find your weapon and your sport of choice, if you will, and commit to being in the top 1%.
But again, I think it starts with simply reading and writing. That's all for this episode. If you'd like to submit a question, please email a voice recording to officehours.com. Again, that's officehours.com. Or if you prefer to ask on Reddit, just post your question on the Scott Galloway subreddit, and we might feature it in an upcoming episode.
