Joe: A company with just 10,000 employees is serving 700 million customers. Meanwhile, a traditional bank needs over 200,000 employees to serve only 67 million. Lewis: Hold on, how is that even possible? That math doesn't seem to add up. That’s a difference of orders of magnitude in efficiency. Are you sure about those numbers? Joe: I am. And that gap is the key. This isn't just about a company being more efficient—it's about a different species of company altogether. And it's a species that is coming for every single industry. Lewis: Wow. Okay, you have my attention. What are we talking about here? Joe: This is the central idea in the book Competing in the Age of AI: Strategy and Leadership When Algorithms and Networks Run the World, by two Harvard Business School professors, Marco Iansiti and Karim R. Lakhani. What's fascinating is they've been teaching and researching this for nearly a decade, and their work on Moderna's rapid vaccine development during the pandemic became a real-time validation of their theories on AI-driven speed and scale. Lewis: So this isn't just abstract theory; it played out in one of the most critical moments in recent history. Joe: Exactly. They argue that AI isn't just another tool for businesses to use. It's a fundamental shift in the very operating system of a company. It’s changing the DNA of what a firm is.

Lewis: Okay, "changing the DNA of a firm" sounds huge. But haven't companies always used technology to get better? We went from paper ledgers to Excel spreadsheets. How is this AI shift any different? Joe: That's the perfect question, because it gets to the heart of their argument. The difference is moving from technology as a tool that supports human work to technology as the core of the operation itself. Let’s go back to that first example: Ant Financial versus a traditional bank like Bank of America. Lewis: Right, the one with the impossible-sounding numbers. Joe: Ant Financial, which grew out of Alibaba's Alipay, has this famous "3-1-0" loan model for small businesses. It takes three minutes to apply for a loan online, one second for the AI to make an approval decision, and there is zero human interaction in that entire process. Lewis: One second? A human loan officer can't even open a file in one second. But isn't a human touch essential in finance? What about risk? This sounds like a recipe for disaster. Joe: You'd think so, but this is where the new operating model flips the script. A traditional bank's growth is constrained by human bottlenecks. To process more loans, you need more loan officers. Each officer has a limit to their expertise and the number of applications they can review. Their performance doesn't necessarily get better with each loan they approve. Lewis: Okay, that makes sense. They might get more experienced, but they don't magically get ten times smarter. Joe: Exactly. Ant Financial’s model, however, is built on a digital core. Its AI doesn't just process the loan; it analyzes thousands of data points in real-time—transaction history on Alibaba, supply chain data, customer ratings, even the business's web traffic. It builds a risk profile that is often more accurate than a human's. And here's the kicker: with every loan it processes, it gets more data, which makes the algorithm smarter. It learns and improves at scale. Lewis: So it's like instead of a human pilot, the plane is the pilot, and it's constantly learning from every single flight every other plane in the fleet has ever taken, all at once. Joe: That's a perfect analogy. The authors call this a "digital operating model." It removes the human bottleneck from the critical path of delivering the service. Humans are still essential—they design the system, they set the strategy, they handle the exceptions—but they are no longer the engine. The engine is the code. Lewis: And that's why they can scale so massively. The marginal cost of processing one more loan is virtually zero. It's just a bit more computing power, which, thanks to the cloud, is basically infinite and on-demand. Joe: Precisely. This creates what the authors call a firm with fundamentally different properties of scale, scope, and learning. They can grow bigger, faster. They can expand into new areas, like Ant did from payments to loans to insurance, because the data and architecture are connected. And they learn and improve at a rate traditional companies just can't match. Lewis: That feels like a truly unfair advantage. It’s like one side is fighting with spears and the other has self-guiding drones that get better with every mission. Joe: It is. And that's why the authors argue that simply creating a "digital division" or opening a "Silicon Valley lab" is a fatal mistake for traditional companies. It's like putting a jet engine on a horse-drawn carriage. You have to re-architect the entire vehicle.

Lewis: Okay, so if the 'digital operating model' is the car, what's the engine? What’s actually doing all this learning and decision-making? Joe: That's the perfect question. The authors call it the 'AI Factory.' It's the scalable, systematic engine that powers the whole digital firm. It’s not just one magic algorithm; it's an entire production line for making decisions. Lewis: A factory for decisions. I like that. What are the components? How does it work? Joe: They break it down into four key parts. First, you have the Data Pipeline. This is the foundation. It’s about systematically collecting, cleaning, and integrating vast amounts of data from every possible source—user clicks, sensor data, social media, everything. Lewis: Garbage in, garbage out. So the data has to be good. Joe: Exactly. The second part is Algorithm Development. This is where data scientists build and train models to make predictions. For example, predicting which customers are likely to cancel their subscription, or what the optimal price for a product is right now. Lewis: And these aren't static, right? They're always learning. Joe: Correct, which leads to the third, and maybe most important, component: the Experimentation Platform. This is where the factory gets really powerful. Companies like Netflix or Google don't just guess if a new feature is good. They test it. They run thousands of A/B tests constantly. Lewis: I've heard about this with Netflix. They don't just have one poster for a show, right? Joe: Not even close. They'll show you a poster for Stranger Things with the kids on their bikes because you watch adventure movies, but they'll show me one with the spooky monster because I watch horror. They test which image gets which type of person to click 'play.' One of their execs once said there are "33 million different versions of Netflix," each one personalized by the AI factory. Lewis: That's incredible, but it also sounds a bit creepy. Are we all just lab rats in Netflix's giant experiment? Joe: The book definitely dives into that ethical gray area. And that's where the fourth component comes in: the Software, Connectivity, and Infrastructure. This is the physical and digital backbone that connects everything, allowing these automated decisions and experiments to be deployed instantly and at scale. Joe: To show you how powerful this "factory" concept is, the authors tell the story of 'The Next Rembrandt.' Lewis: The what? Joe: In 2016, a team used AI to create a brand-new painting in the style of Rembrandt. They fed an AI every single one of his known works, which the AI analyzed for everything—brushstroke technique, subject matter, use of light. The AI factory then generated a completely new portrait, which was then 3D-printed with layers of ink to mimic his brushwork. Lewis: No way. And was it any good? Joe: It was so good that when it was unveiled, many art experts were stunned. It was a perfect imitation. The AI factory had, in essence, reverse-engineered and replicated the core components of a genius's style. It shows that this model isn't just for business decisions; it can be applied to creativity, science, and more. Lewis: So the AI Factory is a system for deconstructing any process, whether it's selling a product or painting a masterpiece, into data, and then using that data to automate, predict, and improve. Joe: You've got it. It's an engine for turning data into action, and it's the reason these new firms can operate on a completely different level.

Joe: And that's where things get really interesting, and frankly, a little scary. What happens when these AI-driven companies 'collide' with traditional ones? Lewis: A 'strategic collision.' It sounds dramatic. Joe: It is. The book uses the classic story of Nokia versus the iPhone. In the mid-2000s, Nokia was the undisputed king of mobile phones. They were innovative, they had great hardware, they had a massive market share. Lewis: I remember. Everyone had a Nokia. They were indestructible. Joe: Exactly. But Nokia was fundamentally a hardware company. Each phone was a separate project, running on a fragmented and clunky operating system called Symbian. They were building individual products. Then, in 2007, Apple launched the iPhone. Apple wasn't just selling a phone; it was selling a platform. It was selling iOS and the App Store. Lewis: It was selling a unified system. Joe: A unified system that was, in effect, an AI factory for apps. They created a consistent platform that made it easy for millions of developers to build and deploy software. This created a powerful network effect—more users attracted more developers, which created more apps, which attracted more users. Nokia's app store, Ovi, was a ghost town in comparison. Lewis: Because it was too hard for developers to build for dozens of different Nokia phones. Joe: Precisely. Nokia was fighting a product war, while Apple and Google's Android were fighting a platform war. Nokia's traditional operating model, with its siloed product teams, couldn't compete with the scale, scope, and learning of a digital operating model. And the result was a complete collapse. Lewis: And the book argues this pattern is repeating everywhere? Joe: Everywhere. Look at travel: Airbnb versus Marriott. Marriott has a hundred years of experience building hotels. Airbnb has an AI factory that connects millions of hosts with millions of guests. It owns no rooms, yet it has more listings than the top five hotel chains combined. It's a collision of operating models. Lewis: This 'winner-take-all' dynamic sounds brutal. The book was praised for its insights, but some critics argue it underplays the resilience of traditional companies. Can't Marriott just build its own AI factory? Joe: They're trying! But the book points to something called "architectural inertia." It's incredibly difficult for a massive, established company to fundamentally re-architect its entire organization. It’s not about buying technology; it's about changing culture, processes, and power structures that have been in place for decades. It’s like asking a battleship to turn on a dime like a speedboat. Jeff Bezos forced this change at Amazon in 2002 with his famous mandate that all teams must communicate through service interfaces, a move that laid the groundwork for AWS and Amazon's AI dominance. But that kind of top-down, painful transformation is rare. Lewis: So this creates a new 'meta' for business, like in a video game when a patch changes all the rules and a new winning strategy emerges. Joe: That's the perfect term for it. The new meta is defined by network effects and data control. But it also comes with a dark side, which the book is very clear about. These same frictionless, scalable systems can amplify bias, spread misinformation, and create massive societal problems. Lewis: Right, like the stories of anti-vax propaganda on Facebook or Amazon's hiring algorithm that was biased against women. The AI factory doesn't have a conscience. Joe: It doesn't. It just optimizes for its goal, whether that's engagement or hiring efficiency. And if the data it's fed is biased, it becomes a bias-amplification factory.

Lewis: So, what's the big takeaway here? Are we all doomed to work for one of five giant AI companies that know everything about us? It feels a bit dystopian. Joe: It could be, and the authors don't shy away from that. But their ultimate message is actually a call to action. They call it a "leadership mandate." The technology is here, and it's creating this new 'meta' where the rules are different. The real shortage isn't AI; it's wisdom. Lewis: Wisdom. What do they mean by that? Joe: They mean that leaders of these firms—and leaders in traditional firms trying to transform—can no longer just focus on optimizing for shareholder value. The scale of their impact is too vast. A decision at Facebook can influence an election. A change in Amazon's algorithm can bankrupt thousands of small businesses. The performance of these digital firms is now intrinsically linked to the health of the vast economic and social networks they depend on. Lewis: So they have a responsibility that goes way beyond their own bottom line. Joe: A profound one. The authors argue that leaders must now think like "keystone" species in an ecosystem. Their job is to ensure the health of the entire network—their employees, their customers, their partners, and the broader community. They have to proactively address the ethical challenges of bias, privacy, and fairness, not as a PR issue, but as a core strategic imperative. Lewis: Because if the network collapses, their business collapses with it. Joe: Exactly. The authors ask us to think about how our own organizations are architected. Are they built for the last century, with siloed departments and human bottlenecks? Or are they being re-architected for this one, with integrated data and agile teams? Lewis: That's a powerful question. It makes you look at your own workplace differently. We'd love to hear what you all think. Drop us a comment on our socials—what's one process at your job that feels like it's still stuck in the pre-AI era? Is there a human bottleneck that drives everyone crazy? Joe: It’s a challenge for everyone, from the CEO to the intern. And navigating it requires a new kind of collective wisdom. Lewis: This is Aibrary, signing off.