Dive Deeper
Transcript
Episode Overview
The end of the billable hour will come from law firms' own clients, argues Bjarne Tellmann, who spent 30 years in senior legal roles at Coca-Cola, Pearson, Haleon, and Aramco before writing Law in the Era of AI (Wiley, 2026). As companies rebuild themselves as AI factories, they consume legal services at a speed and scale the human pyramid model cannot serve, and they will hire whatever serves them instead.
Tellmann's favorite parallel is Nokia. The year the iPhone launched, Fortune put the cell phone king on its cover, and the first iPhone was arguably the inferior handset: expensive, no keyboard, a mediocre screen.
Nokia missed that the job to be done had shifted from the phone to the apps.
In this conversation with GC AI co-founder and CEO Cecilia Ziniti, Tellmann applies that lens to Big Law, walks through the 40% cost restructuring he led at Pearson, and names the third engine of in-house value most legal leaders have yet to build.
About Bjarne Tellmann
Bjarne Tellmann is CEO of FjordStream Advisors and a Senior Visiting Fellow at the London School of Economics. Over three decades in-house, he was founding General Counsel and an Executive Committee member at Haleon, the FTSE 20 consumer health company spun out of GSK, Chief Legal Officer and General Counsel of Pearson, Senior Vice President and Deputy General Counsel at Aramco, and Associate General Counsel at The Coca-Cola Company, where he spent 13 years across London, Vienna, Athens, Tokyo, and Atlanta.
He is the author of Building an Outstanding Legal Team (Globe Law and Business, 2017), the book fellow CZ and Friends guest Chuck Kable recommends to first-time GCs, and Law in the Era of AI: Clients, Firms, and the Future of the Legal Industry (Wiley, 2026). His honors include General Counsel of the Year at the British Legal Awards, the Burton Award, and the Chambers GC Influencers Global 100.
Before law school at the University of Chicago, he was a professional actor in Norway.
Key Takeaways
Clients, not startups, will end the billable hour. AI-factory companies decide at a speed the hourly pyramid model can't service.
The AI factory is structural, not a metaphor. Borrowed from Harvard's Iansiti and Lakhani, it means data flows continuously and gets mined to automate decisions at scale, the way Google runs ad auctions with no human auctioneer.
Tellmann cut Pearson's legal costs 40%+ by restructuring, not cutting headcount. He mapped workflows, set a target internal-to-external spend ratio, and shifted work from premium to mid-market firms.
In-house value runs on three engines. Productivity and pragmatism are old news; governance, building the guardrails for AI systems that decide in milliseconds, is the one the AI era demands.
By 2030, three models survive. Bespoke boutiques, PE-corporatized firms with real CEOs and tech budgets, and "Swiss Army knife" shops that wrap consulting and technology around a legal core.
Will AI End the Law Firm's Billable Hour?
Clients will. Tellmann's argument, developed across 30 years in-house and in Law in the Era of AI, is that companies hire law firms to do a job, and the job is shifting.
Once an AI-enabled legal department can get most work done faster and cheaper, clients will fire the old model and hire the new one, the same way iPhone buyers stopped paying for the best handset.
Tellmann puts the diagnosis in one sentence:
"The irony is you won't fail because you're so bad at what you do. You will fail precisely because you are so good at what you do."
That is Clayton Christensen's innovator's dilemma unfolding in legal, Tellmann says. Nokia doubled down on world-class handsets in 2007 because the iPhone looked inferior on every handset metric.
The job to be done had already moved to the app store. Law firms making world-class arguments at \$1,000 an hour face the same trap. The pyramid model Cravath built in the early 1900s relies on a mountain of junior human talent charging hourly rates for input costs, sold to clients with manifest dissatisfaction about the model.
The work product is excellent, and the job it was hired for is disappearing.
What Is an AI Factory, and Why Does It Change What Companies Need From Legal?
An AI factory is a self-reinforcing operating model in which data flows continuously through a company and gets mined to generate insights, automate decisions, and improve performance at scale. Tellmann borrows the term from Competing in the Age of AI by Harvard Business School professors Marco Iansiti and Karim Lakhani, and he means it as a structural description rather than a metaphor. No human auctioneer runs Google's millions of ad auctions, and no dispatcher at Uber matches cars to passengers.
The model has escaped the tech sector. Tellmann points to Walmart, which put technology at the heart of its operating model after COVID shut its stores, combined its digital marketplace with its physical infrastructure, and, he notes, recently passed a \$1 trillion market cap with a price-to-earnings ratio that beats most of the Magnificent Seven.
When your company becomes an AI factory, what it needs from legal changes with it. Tellmann puts the shift in the CMO's words:
"You know those ads we used to run where we'd give you like two versions to consider and you'd have a week to decide what you thought about it? Well, we can now generate 400 of those and we need your answer by tomorrow. You can't just keep doing things in the same way that you've always done them. You need to up your game and you need to create an AI era legal department."
How Do You Cut Legal Costs 40% Without Gutting the Team?
Restructure the model before you reach for headcount. When Tellmann arrived at Pearson as General Counsel in 2014, the legal department had roughly 230 people, spend ran about 70% external and 30% internal, and the external share sat concentrated in a handful of premium firms.
Rather than cutting people to hit a number, he ran a full diagnostic: map every workflow, find the optimal internal-to-external ratio for the industry, then move deliberately toward it. The result was a cost reduction of more than 40%.
The mechanical insight is the one he repeats to clients today:
"Use a Ferrari when you need a Ferrari and use a Honda when you need a Honda. Just by shifting, let's say, 20% of your spend from a premium firm to a mid-market firm, you save 20 to 30% on your hourly rate. You multiply that by however many hours a year, and right there you're cutting a ton of spend."
Panelizing rarely meant firing a firm. Retaining a firm and using it less is a gentle nudge, and the house firm got the message that it would keep competing for the premium work while price decided the rest.
The deliberate, audit-first version of this playbook is the same one GC AI's guide on how to reduce outside counsel spend with AI walks through for in-house teams starting from a higher external ratio than they want.
Underneath the ratios sat a mindset shift. Tellmann ran the department as a business:
"I kind of imagined myself as a CEO of this legal services business that has one client, Pearson. How are we going to do this in a way that actually returns a profit for us."
The profit margin was the budget, and that framing forced strategy questions most legal departments never ask: why do we have a license to exist, what is our license to win, and what data proves it. Departments that report to the board with that discipline track the same numbers covered in GC AI's guide to legal department metrics.
What Did the Restructuring Teach Him About Change Management?
Tell people why. Tellmann says he still has scars from decisions he made at Pearson, and the sharpest lesson was that a strategy communicated as what and how, without why, causes avoidable pain.
"You can set a strategy and you can tell everyone what you're going to do and how you're going to do it. But if you forget to tell people why you're doing it, that is a painful lesson and it's one that I learned."
The goal is understanding, and understanding differs from agreement. People who understand why, even when they disagree, can come along; the ones who buy in become the missionaries who drive the change.
At Pearson, and later at GSK and Haleon, he interviewed every single person in the function, secretaries included, asked what the team did well and what should change, fed the answers into a word cloud, and put the results back to the team: you told me what to change, now who will help me do it?
What Do In-House Teams Do That No Law Firm Can?
Two things, by definition: prevent problems before they exist, and translate legal advice through deep knowledge of the company's risk appetite, culture, and objectives. Tellmann surfaced both by putting a deliberately uncomfortable question to his legal leadership team:
"If McKinsey came to our CEO and basically said, hey, you know what? We could save 30% of legal costs by outsourcing 100% of legal, why should our CEO say no? Because if you can't answer that question, you're just a commodity."
The team came back with the twin engines that carried his strategy for years: productivity and pragmatism. Outside firms add immense value, but they cannot take the work the final mile, the part where advice becomes something the business can execute.
His favorite proof is a Coca-Cola story. One legal team challenged itself to compress a five-to-ten-page agreement for small-store customers into one page of plain English, stripping every clause that had ever drawn pushback.
Signing time dropped by days. Multiply the recovered sales days across a year and the average life of a customer relationship, and the "riskier" one-pager added measurable profit, even pricing in the one-in-a-thousand dispute.
Why Is Governance the Third Engine of In-House Value?
Because AI systems scale across the enterprise and decide in milliseconds, someone accountable has to set the guardrails: when humans stay in the loop, which policies the AI must follow, how decisions get reverse-engineered, and who answers for them. Tellmann argues this governance engine now sits beside productivity and pragmatism, and that ceding it to IT is the biggest mistake a lawyer can make.
"You could leave that to the IT department. And I think that's the big mistake most lawyers make. If you do that, you're going to be in discovery before you know it."
His metaphor comes from home. Tellmann lives in Germany, where stretches of the Autobahn carry no speed limit until rain or ice triggers digital limits, and the police enforce those limits without mercy.
That is what good AI governance looks like. Dynamic guardrails let the business accelerate safely and activate when conditions demand. Ben Heineman scaled the inside counsel revolution for the age of globalization by putting governance at the center; Tellmann's charge to today's GCs is to do the same for the AI era.
What Will Law Firms Look Like in 2030?
Tellmann predicts a shakedown that sorts the industry into three surviving models. First, boutiques doing bespoke, partner-level work for clients prepared to pay for it.
Second, firms corporatized by private equity, with real CEOs, incentive structures that keep partners, and capital to invest in technology and new pricing. Third, "Swiss Army knives," proto Big Four models that put law at the center and wrap consulting, technology, and advisory services around it.
"There'll be boutiques in the same way that there are still tailors in London that do handcrafted suits for \$10,000. There just aren't that many of them. And every firm I talked to thinks they're in the top 10."
For most firms, he says, the current state is paralysis, and the trap is their own excellence at a job clients are ceasing to buy
When the unit being sold gets cheaper, the value has to live above it, which is exactly where Tellmann says in-house teams, running on productivity, pragmatism, and governance, are built to work.
Recommended Reading
Why Judgment Is the New Superpower for In-House Lawyers in the Age of AI: Anirma Gupta, former CLO of Unity, on the judgment layer that survives when AI absorbs research and drafting.
AI for General Counsel: One Operating Layer for Solo GCs to Full Departments: how in-house teams stand up the AI-era operating model Tellmann describes.





