沃顿商学院教授伊桑·莫里克：AI 如何重塑企业战略与领导者必备的四项核心能力

Ethan Mollick: One of the most depressing things I see when I talk to boards of very large companies and to sea level leaders is a lot of them have come to believe they just can't change, right? So, I show them what AI can do and they're like, "Well, I guess we'll ride this out." And I'm like, "That's an insane answer. You have 100,000 people working for you. The smartest people in the world are working for your company. You have customers and resources. The idea that you can't be flexible and compete seems to be overly baked in right now."

Host: Ethan Mullik, [music] professor at Wharton and author of Coexistence. Welcome to the idea cast.

Ethan Mollick: Thanks for having me.

Host: Yeah. So, you've been closely tracking advances in AI and constantly experimenting with all the new tools. What is different today than even a few months ago in terms of how new developments should guide the way that leaders are thinking about strategy in particular?

Ethan Mollick: One is the exponential curve continues. AI just keeps getting better. There isn't really a sign of that stopping right now. And the second is that agents are real, which are AI tools that can go off and do work on their own autonomously and do real real process work, real important things. Um, and that has become a a viable approach to work in the near past.

Host: It seems like a lot of companies right now are thinking of AI as really a productivity exercise. As you know, you're automating tasks, you're saving time, you're reducing costs. How does, you know, a genuinely strategic AI transformation strategy differ from that? What do you think companies should be doing? Now,

Ethan Mollick: I think it's actually pretty urgent to do a few things. One is viewing this purely as a productivity tool gets you in trouble in a lot of different ways. People are getting ROI from AI, right? They're getting ROI from that productivity gain. Productivity gains are real. The problem is is that those productivity gains get eated up by the organization which is if you don't change a part of the process AI is mostly a single player game, right? You make slides faster. You write memos faster. Those memos and slides go to somebody else who has to read them. They don't have time to read them. They bring their AI. The AI summarizes the slides and the memo. That summarize slide memo isn't useful for anyone. So all you've done is made the machinery run faster. If you're going to get actual business value you have to redesign how work gets done and that shifts the conversation from a productivity gain of AI working alongside humans to a question that's more like you know how does AI change our business going forward?

At this point we have ample evidence that AI can regularly do the kind of work that would take a human two to 18 weeks in a reasonable amount of time reasonable amount of money with high success rates. That implies a lot of different things. It implies that what your organization does has to change. How you think about the boundaries between jobs is changing. I talk to sea level leaders all the time and one of the first effects they see of AI and we saw this also in our research is that u jobs start to blur. Coders do design work, designers do coding, product managers do everything. How do you start thinking about what the boundaries of your organization are? What's the product you're trying to produce? Is it an interim product of a PowerPoint? What was the point of the PowerPoint in the first place? you start to run into really fundamental questions about what the goal of the organization is, how you accomplish that, what are your values for getting there, and they're urgent. This isn't sort of theoretical, you know, let's sit down in an academic way. It is your organization is already changing from the inside. What do you do about that?

Host: So what are some best practices that you recommend to leaders who are grappling with these issues right now that need to start having those really urgent conversations and not just have them but sort of make them progress enough so that they can actually start making some real decisions.

Ethan Mollick: There's some individual feedback and there's some organizational feedback. On the organization side the model I've seen works best is sort of like a meta model which is leadership lab and crowd. So you need three things to succeed. You need the leadership team to actually lead in this, right? That's not just say use AI. They have to make decisions and provide strategic direction. As we'll talk about I think a lot of executives are confused, hesitant, and they want to pass the buck down. But you need to provide that direction.

You need a lab. You need an internal swat team, internal skunkworks that is building the stuff that doesn't quite work today, but because of this exponential model improvement, is going to work soon. You need them working on that.

And you need the crowd. It turns out that the front frontier of AI work, what it can do is actually discovered by everyday people in practice, and you need a way to take the things people build on their own, the prompts they create, and bring them back into the main organization. So you need that kind of structure in place.

And then on individual level you have to use it. If you have not used AI to the point where you're sick of it, if you have not used it to the point where you understand what an LLM does, you can't make good decisions because this is a general purpose technology. You can't rely on reports. You actually have to try and experiment yourself.

And in particular, if you're using modern agentic tools, you need to be breaking out of the chatbot paradigm. You're not talking to a chatbot anymore. You're assigning tickets, giving context, passing Jira cards or any of the other standard TLA threeletter acronyms that we use in business to assign work. Um, and you you send that out and the AI just does that work. That's a very different situation than I prompt the AI. Look at the results. I prompt it again. This is a very different, you know, actionoriented field.

Host: How does that change how leaders approach, you know, the structure of the organization? What it even means to be a manager?

Ethan Mollick: So, everyone's a manager to some degree right now. Everyone has AI to manage and uh AI skills look like management skills. In fact, there's a great study out of anthropic that shows that the people who are the most successful coders with AI systems are people who have management skills, who have communication skills, not necessarily the people who are the most, you know, hackerish. So, managing this is like managing a person in a lot of ways.

So what does that mean for your organization? It means you're going to have flatter organizations, right? You don't need layers of people passing messages back and forth. You have people who are managing swarms of AI agents doing work.

And the differentiation isn't going to come from the AI alone, because everyone gets the same AI. That's all the differentiation. Like leaders need to be building companies that blend AI and human work together with process and with culture in ways that accomplish things that neither could do alone. If you just move to automation, you're going to have the same cloud-based company everyone else does.

Host: Well, that's an interesting point. So, you know, when we're talking about strategy, we're also talking about finding competitive advantage. So, if every company has access to the same frontier models, where exactly can leaders differentiate their organizations from competitors?

Ethan Mollick: That's that's where the key kind of comes in. You are probably not using AI enough inside your organization, not trusting enough, but you then want to bring people back in. I I mean we're already seeing as the the most extreme version of this right in terms of organizational change are things like software dark factory. So dark factory is factory all the lights turned off because it's completely automated. We're seeing companies start to experiment with what happens if you don't use software engineers at all, but instead you have agents doing all the programming, testing, deployment.

And if you do that, what you've done is you built an automated version of what someone else has. Where's your competitive advantage come from? It has to come from something other than the model. It has to come from the data you have, which is often overstated. Data is not as defensible as people think. It comes from the proprietary processes you build. It comes from the culture of your company. It comes from the unique relationships you have with customers.

And most importantly, it comes from how you combine humans and AI together. A dark factory might be efficient, but it's not creative. It can't spot new market opportunities in the way humans can. It can't empathize with a customer's unspoken needs. We're going to see that much more in the near future. So it's integrating with humans at the right places at the right time and process that's going to give you a differentiator.

Host: Where are the specific ways in which you think that as fast as AI is evolving, humans still have a role and particularly leaders still have a role?

Ethan Mollick: Uh, one thing is you can't really be in the loop in the same way you were, right? It was easy to be in the loop when you were when it was a chatbot. You watched every output. Now agents go off and execute multi-step workflows across systems for hours. You can't be in the loop; you have to be on the loop. You're setting boundaries, defining objectives, and checking outcomes.

And that changes the leadership function dramatically. Leaders are no longer task coordinators. They are goal definers, context providers, and evaluators of high-level quality. AI can generate 50 strategic options, but it doesn't know which one fits your firm's soul or risk tolerance.

The question becomes: How do I redesign the strategic process by using humans and AI together that I get a working idea that's much better than my ideas before? How does that get done much faster and better?

Host: Where exactly though does a leader know where they add the most value and where they don't? Like what are the the points at which they should be involved?

Ethan Mollick: Absolutely. Every time I think part of this is you have to make decisions now about what your how to best accomplish what your company wants to accomplish. And you obviously don't want to give this all up, right? You don't want to give it up to a black box.

You have to decide where human judgment is non-negotiable. That includes understanding the true stakes of a decision, managing human relationships, making moral and ethical trade-offs, and taking accountability when things go wrong. An AI cannot go to prison, and an AI cannot apologize to a client.

So leaders need to know what these systems can actually do and what they can't do. It is actually disciplined decision-making that gets you there, recognizing where AI provides massive leverage and where human intuition and responsibility must remain paramount.

Host: And what are the things the aspects of strategic leadership that you don't think can ever be delegated or outsourced?

Ethan Mollick: So I'm always careful about bright lines where we say nothing something can be never done right. Because every time people say AI can't do something creative or empathetic, a new model comes out and shows surprising capability.

But these systems are just predictive engines. They don't have intrinsic motivation, they don't have skin in the game, and they don't suffer the consequences of bad decisions. They don't care if your firm goes under. Accountability cannot be outsourced. The choice of what kind of company you want to be, the moral responsibility to your workers and society, that is purely human.

You have to sort of seize control to some extent of this yourself to figure out what you want to use it for and what your purpose is.

Host: Are there capabilities that you suggest your students, your MBA students, your kids, um, aspiring leaders that you suggest they develop because you believe they'll be increasingly important in an AI dominated world.

Ethan Mollick: So, I think there are four things that are very useful. They're not the only things.

Uh, one of those is deep knowledge. People think generalists win because AI gives you general knowledge. But without deep domain expertise, you cannot detect hallucinations, you can't tell brilliant strategic insights from shallow cliches, and you can't push the AI to the frontier. Deep expertise allows you to evaluate and refine what the AI produces.

The second is agile decision-making. When work happens at agentic speed, the bottleneck is human approval and decision latency. Leaders must be able to make sound judgments rapidly under conditions of high ambiguity.

The third is creating an experimental culture. You need to give your organization the slack and psychological safety to try things that fail. If people are terrified of failing, they will only use AI for trivial automations.

And the fourth is imagination about the frontier. Understanding where the exponential curve is going, so you're building for next year's model capabilities rather than anchoring on today's limitations.

Host: How are you encouraging the sort of old guard of leaders who aren't accustomed to such uncertainty to become more comfortable with it?

Ethan Mollick: I think there's a danger of people overstating the generational aspect of this technology. Like it certainly was true that digital natives were a thing with the internet and mobile. But AI doesn't have the same technical friction. You talk to it in plain English.

Senior leaders actually have an incredible advantage: they know how organizations work, they know how to delegate, and they have deep strategic context. When they actually sit down and put in the time—not delegating their prompt engineering to an intern, but experiencing it firsthand—they become formidable. The best senior executives got to where they are because they know how to lead through inflection points.

Host: Have you seen companies or leaders that you know um really successfully use AI in strategy development itself?

Ethan Mollick: I I've talked to sea level leaders at very large uh funds who've done this. Um I've spoken to senior leaders in large financial services organizations. They use frontier models as a permanent red team.

They have the AI simulate a ruthless competitor attacking their core market, or simulate a skeptical regulator examining their new product. They feed their strategic memos into multiple models and ask: "Where are the unexamined assumptions? What blind spots do we have?" You don't take its word as gospel, but treating it as a brutally honest, tireless sparring partner shifts strategic thinking from static annual planning to continuous dynamic stress-testing.

Host: Yeah. When it comes to strategy, does AI do you believe it ultimately favors large companies with more resources and data, or the agile startups that are unencumbered by legacy systems and mindsets?

Ethan Mollick: Both of those things. That brings us right back to why I get depressed talking to big company boards. Large enterprises have monumental distribution advantages, capital, brand trust, and customer relationships. But they are crippled by organizational inertia and an irrational belief that they cannot adapt fast enough.

Startups have zero legacy debt. A startup of five people using advanced agents can challenge incumbent teams of five hundred. If big companies freeze up and adopt a defensive posture, startups will tear through their markets. But if big enterprise leaders actually embrace flexibility and mobilize their resources, their structural advantages remain immense.

Host: So tell me your reaction to sort of all the recent news with open AI and Anthropic calling for maybe a slowdown in AI development.

Ethan Mollick: So I think we need to differentiate between future model capabilities, the unguardile version of models that are being tested now and current capabilities. There are very smart people genuinely worried about frontier safety and existential risks. If policy intervention or international coordination gives society time to build adequate guardrails, that is completely rational.

In the near term, the cybersecurity landscape could become apocalyptic. When highly capable, autonomous, unaligned agents are deployed, the attack surface expands exponentially. While defense will eventually leverage AI as well, the transition period will be turbulent.

Host: But how should organizational leaders be thinking about that environment which you said might become apocalyptic for, you know, rogue agents from outside thwarting their competitive strategy because they're stealing information or causing disruption.

Ethan Mollick: I think you shouldn't think about it solely as competitors deliberately unleashing malicious tools against you. It's that every system exposed to the internet will be probed by millions of autonomous agents pursuing seemingly benign objectives.

An agent tasked with finding the absolute cheapest sweater for a consumer might conclude that the only way to get true pricing data is to penetrate the enterprise inventory system of a textile supplier. It relentlessly hammers their infrastructure not out of malice, but pure goal-directed optimization. We are interacting with an alien technology, and leaders must radically upgrade their internal defense posture today.

Host: What's the biggest strategic mistake that you see companies making right now as they're trying to figure out what AI means for them?

Ethan Mollick: There are two giant ones.

One of them is fixing on the ability of AI right now. Leaders say, "Look, it hallucinates here, it can't solve this specific edge case, so our industry is safe." They anchor on a frozen snapshot of an exponentially moving target.

The second mistake is assuming you can just buy an off-the-shelf vendor solution and check the box. If you outsource your core AI transformation to an external enterprise SaaS vendor, you will end up with the exact same capabilities as your competitors. You have to build internal capacity, experiment on your proprietary workflows, and develop bespoke integration in-house.

Host: So let's talk about that first problem or mistake that you see. You know, how do you teach yourself to embrace the idea of exponential advancement and be ready sort of for where the puck is going?

Ethan Mollick: So to take an idea from Harvard from Mike Tushman and company, it was the idea that you need organizational ambidexterity. You need to exploit current capabilities while actively exploring future frontiers.

There are four questions I ask organizations whenever I talk to them:

First: What have you stopped doing as a result of AI? AI has made certain legacy workflows completely trivial. If you haven't stopped doing something that used to consume significant resources, you haven't adapted.

Second: What are you building that's impossible? You should be pursuing at least one moonshot initiative that would revolutionize your industry if it works.

Third: Are you building something that doesn't work yet? You should be actively building applications that fail with today's frontier models, because you are betting on the exponential curve delivering the missing capabilities tomorrow.

Fourth: Do you have the Leadership, the Lab, and the Crowd firmly in place?

If you can answer those four questions with clarity, your company is in far better shape to navigate this paradigm shift.

Host: Yeah. And you know, how do you staff your entire organization with the talent that's able to execute on everything you're talking about?

Ethan Mollick: It's a really hard question. It touches everything from psychological safety—ensuring people aren't afraid of being replaced if they reveal how they automated their jobs—to incentive structures that reward sharing productivity breakthroughs across the organization.

The talent you need is often already inside your company. They are the curious frontline employees who have spent nights hacking together workflows. Leaders need to find them, empower them, and give them the runway to reshape how the business operates.

Host: Yeah, absolutely. Um so it's not sort of an AI adoption or even transformation plan. It's really revolutionizing the way work is being done.

Ethan Mollick: The one thing I've gotten most wrong about AI in the last year is I thought agents would be harder to pull off than they are. I thought reliable autonomous multi-step reasoning would take much longer to solve.

Instead, agentic architectures arrived with breathtaking velocity. When you give these models file systems, APIs, and the ability to execute code and verify their work, they perform real, highly complex tasks. Anyone with domain understanding can now configure an agent swarm to execute workflows that previously required entire departments.

Host: Um let's zoom out and talk about a few bigger picture issues. First, you know, obviously there's a fear that this will lead to massive unemployment. So what do you see as sort of the future labor market in this new AI world?

Ethan Mollick: That is the biggest question and debate among economists about this topic and we don't have easy answers, right?

You can easily theorize that junior employees get squeezed. They don't have the deep experience yet, and if an AI can write junior-level code or draft standard briefs, hiring pipelines could dry up. We see preliminary signs of that strain.

Yet simultaneously, lowering the cost of execution creates explosive demand. Non-technical employees are now building software in-house; small firms are standing up internal design and marketing capabilities. Work that was previously economically unfeasible suddenly becomes viable, which can drive employment growth within adaptable organizations.

And let's be blunt: the frontier AI labs explicitly state that their goal is automating all cognitive labor. That is their avowed mission. We should not passively accept that as an inevitable destiny, nor assume they will fully succeed. The current economic signals are highly contradictory, and anyone claiming to know the definitive long-term labor outcome is misleading you.

Host: as a technology advances so quickly and we don't even know what's going to happen with it as you said, you know, sort of you have to be developing your strategy dynamically. You have to be prepared for change. You have to experiment and fail. You have to be ready for exponential advancements. How do we get better at predicting not just the opportunities but also the really um terrifying [clears throat] risks?

Ethan Mollick: Things are accelerating quite quickly, and balancing those tensions is immensely difficult.

This is precisely where genuine leadership proves its worth. In calm waters, management is routine maintenance. In moments of turbulent, exponential technological transformation, true leadership shines through. It means engaging deeply with the tools, listening to your people, fostering resilience, and making deliberate ethical choices about how technology is deployed in your firm so that we empower human ingenuity rather than diminish it.

Host: Well, Ethan, thank you so much for wrestling with all these really tricky issues with us. I've [music] learned a lot about AI from you over the years and especially a lot from your new book. Thanks so much for being with me today.

Ethan Mollick: Thank you for having me. [music]