In a Q&A, Bjarne P. Tellmann expands on his book Law in the Era of AI (Wiley 2026).
The Practice: Why tell the story of AI’s impact through the evolution of the corporation and the general counsel? What does a historical lens grounded in business and globalization reveal that a tech-first account or even a legal profession-first account misses?
Bjarne P. Tellmann: The legal profession is one of the few ones in the world where we have this strange relationship to clients in which we’re relentlessly focused on representing clients to the highest possible standards. And yet we have a remarkable lack of curiosity about what’s going on inside clients’ businesses. I spent 30 years working with CEOs of large companies, and they do nothing but worry all day about what’s happening with their customers.
The corporate lens is going to tell you a lot about where the industry is heading. Corporations are about 75 percent of total legal spend, according to the Bureau of Labor Statistics, and that chunk of spend is going to drive the pace and the structure of supply. If you want to understand what AI is going to do to law, you have to understand what it’s doing to corporate clients and how it’s changing the way that they buy, the way they organize, the way they consume legal services.
I think the historical lens is also useful because it shows that legal institutions historically have transformed as a result of forces that emanate outside of the profession. Taking that historical view prevents us from mistaking current structures as something that will remain permanent.
Zeroing in on the GC and its historical evolution is important because that role sits at the intersection of business, law, and institutional change. If you look at it, the history of the general counsel is actually in some ways the history of how companies have adapted to risk and opportunity throughout history.
The Practice: A through line across your four GC eras is the “twin dynamics of risk and efficiency.” Is GC 4.0 a continuation of the “more for less” pressure that defined GC 3.0, or is generative AI something categorically different?
Tellmann: In many ways it’s both. The more-for-less dynamic will continue, but I think that AI is changing how lawyers address the more-for-less challenge.
The primary focus of the GC 3.0 era was how to optimize human service delivery. How do you streamline workflows? How do you manage costs? How do you deploy people in the right way around your service model? In 4.0, we now have a technology that actually participates directly in the cognitive work and is taking on ever more of that work. That’s categorically different. It introduces a whole range of fascinating new dynamics, and the first one is that the economics change. The more work that AI takes on, the more the key question for GCs moves from how do you make human lawyers more efficient to determining where humans fit in or if humans are needed at all, and what the split should be between human and artificial intelligence and how do you optimize that mix?
The second interesting dynamic is how the opportunities and risks change. In GC 4.0, this balance is changing because the corporation itself is changing at a breathtaking pace. It represents a fundamental reset in how companies establish themselves and how they derive profit. A digital business can scale at near-zero marginal cost; it can scope, meaning expand the industries and business areas where it is active, at relatively low cost; and it can continuously improve and optimize.
Everyone is leaning in, but it also changes the risk equation. As you put these systems at the heart of your operating model, they start taking autonomous decisions at machine speed and scale. All of that needs to be managed and governed effectively, and GCs will be at the heart of that equation.
We’re talking about an industrial revolution, and the risk and efficiency dynamic is going to play a big, big role.
All of this is also accelerating. This isn’t just a normal investment cycle—AI is a general purpose technology that is having far-reaching effects. This year alone, globally, $2.6 trillion is being spent on AI and AI-related infrastructure. According to Gartner, worldwide AI spending is expected to reach approximately $2.6 trillion in 2026, up nearly 47 percent from 2025. By 2027, spending is projected to approach $3.5 trillion, representing almost a doubling in just two years and a compound annual growth rate of 41 percent.
In the United States, Amazon, Microsoft, Alphabet, and Meta alone are spending $725 billion on AI and AI-related investments this year. That’s slightly more than 2 percent of U.S. GDP. That exceeds in percentage terms what the U.S. spent on railroads in the 1850s. It’s about 10 times what the U.S. spent on the Apollo mission to the moon. We’re not really talking about a tech upgrade. We’re talking about an industrial revolution, and the risk and efficiency dynamic is going to play a big, big role.
The last thing to say about this is also what’s interesting about the risk-efficiency dynamic in the AI era: they’re emanating from the same source, and that makes it super hard for GCs to persuade leaders to slow down. You’re not going to win that one by saying, “We need to put humans in the loop everywhere and slow the whole thing down.” Instead, the formula has to be one of structured velocity. Take the Autobahn in Germany. It’s a great analogy because there’s no speed limit for large chunks of that road system. The only reason it works is because there are rules and there are norms and the infrastructure is designed for high velocity at a sustainable pace and level. And that’s what GCs need to do in the era of AI.
The Practice: You build on the idea of the “AI factory.” What does it actually mean for a company to become one?
Tellmann: The “AI factory” comes from Harvard Business School professors Marco Iansiti and Karim Lakhani’s Competing in the Age of AI. What they predicted has played out over the last few years not just in digital companies but also in traditional ones. An AI factory is effectively a system where data comes into the company’s operating model, where it gets filtered and organized in ways that allow for AI and agentic systems to read that data and generate recommendations, insights, and ultimately outputs that translate into business outcomes. Increasingly those insights and those actions are being generated autonomously.
Underneath it all, as data gets generated, it gets fed back into the engine where it continuously improves just by operating. Tesla is an example of that: that every vehicle on the road learns from every other Tesla on the road, and the whole system gets better just by operating every day.
In my book, I talk about a number of examples of nondigital companies that are moving in that direction. John Deere, for instance, is a tractor and combine company that was founded in 1837. For most of its history, they had a pretty straightforward model of selling agricultural equipment and manufacturing equipment. Increasingly what they’ve done is tear a page out of the Tesla playbook and connect all of the equipment they’re selling to software, data, and intelligence. Think about a modern Deere tractor being effectively a sensor on wheels. It has cameras, radar; it has GPS and LiDAR; and it’s embedded into a computing architecture. All of those sensors generate enormous amounts of information about the machine itself but also about the conditions in the field, what’s happening in every square foot or, in some cases, even narrower areas of the field. All of that data flows into Deere’s cloud-based operations center where it’s analyzed and combined with weather and other data, and then recommendations are made that flow back into a farmer’s app that allows the farmer and ultimately the machine to have a measurable impact on efficiency.
For instance, John Deere has a seed and spray technology that reduced herbicide use by an average of 59 percent across a million acres of farmland, saving eight million gallons of herbicide mix. If you think about that for a moment, that’s not AI as a chatbot. That’s AI changing the fundamental economics of an industry. It’s also changing how John Deere derives profit. Rather than just selling one-off tractors and saying goodbye to the customer after they sell it, they’re building a relationship and software as a service that brings in recurring revenue and generates a flywheel effect of continuous improvement. The more machines they have wired up to this, the more data they generate, the better the models and predictions and the better the outcomes are, and that attracts more users to John Deere.
The Practice: Now how does this spill over into the law firm space? You focus on how clients will drive firms to change—for example, the billable hour will fall or adapt—and new firm models must emerge. What will these new models look like?
Tellmann: What you’re going to see is a demand for a much more diverse range of models and providers. If you go back to the early 1900s when Cravath was the first mover in professionalizing law firms, you effectively had a one-stop-shop model all the way until basically the beginning of globalization in 1989. That is now fragmented, and that’s structural in nature. GCs are going to require a plethora of different providers. I think what you’re going to see is the law firm model evolve. You’re going to have a range of different delivery structures that meet those different needs.
The billable hour, which has been the anchor for the law firm model for decades and decades, is increasingly under pressure because AI attacks the very basis and assumption of that model. As that changes, that forces law firms to think much more strategically about what they do and how they do it.
It’s very hard to predict exactly what models are going to emerge, but based on what we see in the market today, some firms will become boutiques. They will recognize that the real value that they generate is based on the knowledge and wisdom and insights of their senior partners, and they will coalesce around models that scale that. You might have an inverted pyramid where you scale out the number of very senior people and you charge based on the value of the output, not the cost of the input. Then everything below those senior partners increasingly gets taken over by technology, outsourced providers, and maybe the client themselves.
Another structure I see unfolding is what I call the Swiss Army knife: firms that basically tear a page out of the Big Four and provide legal services at the core rather than accounting, but then around that, they also provide a range of other services that clients need these days, such as technology consulting, risk advisory, etc.
You could also imagine corporate models of law firms taking off. You could see capital infusion accelerating technology investment, new forms of billing, new strategies, a top-down CEO-led decision-making hierarchy inside those firms, different ways for partners to draw profit rather than just through partner shares. And then there are capital consolidators: PE-backed outfits that target different parts of the legal services industry—not necessarily just legal advisory but all the stuff that is needed to support law firms. That could be a complement to the boutique. The logic there is very similar to what they say happened in the gold rush: the people that made the money were the ones selling the shovels, not the ones mining the gold.
All of these models will have different shapes: diamond and hourglass and obelisk and inverted pyramid. What’s interesting is that the shape of the organization will follow the logic of its strategy. The core insight there is that law firms have to get serious about strategy. They can no longer see strategy as just a U.N. compact between partners and how do we best accommodate everybody. That’s not going to work. They’re going to need to sharpen that and really think about what it is they do that is unique and why they have a right to win in the market, kind of like every company.
The Practice: This moment is messy, as the legal profession tries to integrate AI but hasn’t quite figured out how to fully realize its efficiencies and benefits, and it’s hard to document transformation as it’s happening. What are lawyers still struggling with, and what does the legal profession look like to you in 10, 20, or even 50 years?
Tellmann: It is messy, and that’s one of the most interesting things about being active in this space right now. There’s some very sophisticated first movers, both on the client side and on the firm side. In some cases there’s a handful of firms that are quite far ahead, and they’re far ahead in some cases partly just because of the lucky nature of the market that they serve, which allows them to be structured in a way that wins in the AI era. In other cases, they’ve made conscious decisions that have been adopted by the partnership. It’s the same on client side. You see some clients, in many cases, just because they’re in a digital company or a company that is committed to an AI factory model really moving very quickly ahead.
And then you have a long, long trail of providers and of clients who are just completely confused. That’s going to shake up over the next few years. There’s going to be some rapid advancements: the reshaping of institutions, governance systems, models, professional rules, legal education. Everything needs to coalesce, and it’s all very, very messy right now.
I do think that human judgment, wisdom, and responsibility are things that will continue to be prized on the human side.
If there’s one thing coming out of it that I see on the client side, it’s general counsel struggling with governance and trying to think about how to insert themselves into that conversation of how do they optimize their own workflows to leverage this.
On the firm side, it’s the strategy we talked about, and it’s how do they derive efficiency out of this new AI era. For many of them, they’re putting the systems in place, but they don’t trust the systems, which means they’re paying the fees on the technology side, but they also have juniors that have to check everything, and they have clients asking, “Where are my savings?” That means they’re finding costs are going up, not down.
I struggle to really picture what the world will look like in 20 years, let alone 50 years. I struggle to really think about what institutions will exist, what organizations will look like, what customers will expect. The only thing I can say that’s certain is that society is still going to need rules, it’s going to need rights, it’s going to need legitimacy in the rule of law. It’s going to need dispute resolution and predictability for commerce and for human society to flourish. That means those foundations will remain.
The other thing that will remain—and here is where I differ from some prognosticators—I do think that human judgment, wisdom, and responsibility are things that will continue to be prized on the human side. I believe humans trust humans. They don’t trust necessarily machines. There’s going to be a role for wise humans and humans that exercise that level of judgment. How they get that judgment and wisdom is a whole ’nother chapter that if we had more time, we could talk about.
Related content
Lead Article
Related content
Law in the Era of AI
Bjarne Philip Tellmann advises leading law firms and in-house departments on organizational excellence and digital transformation in the legal industry through his consultancy, FjordStream Advisors. Previously, he served as founding general counsel and member of the executive committee at Haleon, a FTSE 20 company and the world’s largest consumer health business. At Haleon, he led the legal and compliance work associated with the company’s demerger from GSK, among the largest and most complex corporate separations in recent U.K. corporate history. Previously, as chief legal officer and general counsel at Pearson, a FTSE 100 global education company, Tellmann redesigned and digitized the legal department, reducing costs by more than 40% while improving service quality. He has also held senior leadership roles across Europe, Asia, the Middle East, and the U.S. at Coca-Cola, Aramco, Kimberly-Clark, and the law firms Sullivan & Cromwell LLP and White & Case LLP.
Image Credit: Shutterstock // Kozak_photo
