---
title: "Who Owns the Legal AI Harness?"
episode_number: 5
canonical_url: https://enterprisealignedai.com/episodes/who-owns-the-legal-ai-harness
md_url: https://enterprisealignedai.com/episodes/who-owns-the-legal-ai-harness.md
last_updated: 2026-07-22
duration: "48 min"
youtube_url: https://www.youtube.com/watch?v=voxL8usOV-A
spotify_url: https://open.spotify.com/episode/6WqVc22pg349vEbDSodF0D
apple_url: https://podcasts.apple.com/us/podcast/who-owns-the-legal-ai-harness-baker-mckenzie-legora/id1890781384?i=1000778302396
guests: ["Danielle Benecke", "Max Junestrand"]
tags: ["Legal AI", "Harnesses", "Build vs Buy", "Agentic AI", "Compliance", "Pricing"]
---

# Ep. 5: Who Owns the Legal AI Harness?

**Summary:** Host Aparna Sinha sits down with Danielle Benecke, who founded and leads Baker McKenzie's Applied AI practice, and Max Junestrand, CEO and co-founder of Legora, on the question the legal industry has quietly moved to. The conversation starts with what a harness is and ends with who should own one. Danielle argues the real opportunity is not accelerating the work lawyers already do but building services that could not exist before, and that the third wave of AI in legal will be runtime compliance steering for enterprises running agents across the business. Max explains what a legal harness decides that a general one cannot: which documents the model reads, what counts as authoritative between a court case and an internal precedent, which models run where, and where sub-agents are allowed. Along the way they get into why generating legal analysis is now cheap while judgment stays scarce, why consumption pricing arrived in legal AI, why one general counsel's AI-assisted first pass made the work more expensive rather than cheaper, and how a global firm decides which capabilities to buy and which to keep inside.
## Why listen

Most AI conversations in the enterprise are still about which tool to roll out. This one is about what happens after that question is settled: which parts of the AI stack a firm should own, and which it should rent. Danielle's generalizable-to-custom and strategic-criticality axes give you a way to make that call. Max gives the clearest working definition of a harness available anywhere, using Claude Code and Cursor as the reference points. Both are operators describing decisions they have already made, with the commercial consequences attached.

## Key takeaways

- The industry question has moved from which tool to roll out to which harnesses a firm should own rather than rent
- A harness decides what the model reads, what is authoritative between a case and an internal precedent, which models run where, and whether sub-agents are allowed; the simplest harness is a folder of markdown files
- Own the harness where the work turns on your expert steering, internal playbooks, and know-how; rent it where the capability is generalized and not strategically critical
- Enterprises run multiple harnesses at once: direct model access, productivity-layer harnesses, vendor tooling, and their own bespoke systems per service line
- Generating content and legal analysis is cheap now; judgment and accountability are scarce, so the goal is steering agents upstream rather than reviewing more output downstream
- The three waves in legal AI: tools for lawyers, then AI-native legal and compliance businesses, then compliance systems that steer agents and business behavior in real time
- The bigger opportunity is complex work humans could not do before, not accelerating the lower-level work the market is focused on
- Marketing was first to roll out AI at many enterprises, moving legal review from tens of pieces of content to tens of thousands across the long tail of countries
- Expertise now sits upstream in designing the system; the better the design, the less downstream validation is needed, which changes what a firm sells
- Consumption pricing came to legal AI because these businesses do not have software margins, while law firms move the other way toward fixed fees to absorb that uncertainty for clients
- One general counsel asked firms to quote both validating her team's AI-generated first pass and doing the deal end to end; validating the AI first pass came back more expensive
- A more expensive model can be the cheaper choice, because token efficiency varies and a cheaper model can use more of them

## Timestamps

- 0:00: Legal AI: Whose Harness Is It Anyway?
- 0:33: Meet the Guests: a Global Law Firm and a $5.6B Legal AI Startup
- 1:45: Why a Law Firm Started Building Its Own AI
- 3:10: Legora: How This High-Growth AI Startup Was Founded
- 4:57: Legora's Vision: Making Law More Accessible and Creative
- 6:57: Baker McKenzie on the New Legal Work AI Enables
- 9:24: Baker McKenzie: The Three Waves of AI in Law
- 10:57: Legora's "Legal Engineer": A Lawyer Who Builds AI
- 12:50: Legora's Real Moat: Deep Enterprise Integration
- 17:42: What Legora's Legal Harness Does That Claude Code Can't
- 21:05: Baker McKenzie: When to Own vs Rent Your AI Harness
- 22:26: Baker McKenzie: Use Human Judgment Upstream to Design the System
- 25:05: Baker McKenzie: From Billable Hours to Fixed Fees
- 26:25: Legora: Why It Moved to Consumption-Based Pricing
- 28:01: How Legora Scaled From $1M to $100M in Under 18 Months
- 34:03: Baker McKenzie: Same-Day Legal Work in a Cyber Attack
- 38:15: When a Local Lawyer Beats a Frontier AI Model
- 39:46: When Using AI Costs the Client More
- 40:37: Build vs Buy for Legal AI
- 44:12: How Legora Stands Out in a Crowded Vendor Market
- 47:10: The Future of Law and Lawyers

## Clips

Short segments published from this episode. Each is citable on its own.

- **Legora: Making Law More Accessible and Creative**
  - URL: https://www.youtube.com/watch?v=SSDURRKRfvQ
  - Speaker: Max Junestrand
  - Occurs at: 4:57
  - Duration: 0:33
- **Three Waves of Legal AI: Compliance Is Next**
  - URL: https://www.youtube.com/watch?v=122-zVuVrKo
  - Speaker: Danielle Benecke
  - Occurs at: 9:24
  - Duration: 0:46
- **What Legora's Legal Harness Does That Claude Code Can't**
  - URL: https://www.youtube.com/watch?v=ASbV3-js0Ww
  - Speaker: Max Junestrand
  - Occurs at: 17:42
  - Duration: 1:02
- **Legal AI: Use Human Judgment Upstream to Design the System**
  - URL: https://www.youtube.com/watch?v=ZdYKJ-VVF-4
  - Speaker: Danielle Benecke
  - Occurs at: 22:26
  - Duration: 0:18
- **Baker McKenzie: Same-Day Legal Work in a Cyber Attack**
  - URL: https://www.youtube.com/watch?v=9rmd1r-V_2Q
  - Speaker: Danielle Benecke
  - Occurs at: 34:03
  - Duration: 0:53
- **From Billable Hours to Fixed Fees**
  - URL: https://www.youtube.com/watch?v=gGlYAohxJxM
  - Speaker: Danielle Benecke
  - Occurs at: 25:05
  - Duration: 0:41

Danielle Benecke founded and leads Baker McKenzie's Applied AI practice. Max Junestrand is CEO and co-founder of Legora. They come at enterprise AI from opposite sides of the table, one buying and building inside a global law firm, the other selling the platform. They agree on the question that matters now: which parts of a firm's AI it should own, and which it should rent.

## The question underneath the tooling

Danielle describes the shift in how her clients think about their AI stack:

> "They're starting to think about what harnesses they should own rather than rent."

Her framework uses two axes. One runs from generalizable to custom, the other from strategically critical to not. Generalized capability that is not strategic can safely run on a platform you do not own. Work that turns on your expert steering, your internal playbooks, and your own know-how should stay in-house.

## What a harness is

Max gives the working definition, using coding tools as the reference:

> "Claude Code is optimized for coding. It goes back to how do you let the model read the documents, make sense of documents, make sense of context? What's authoritative between that case over there and some internal precedent that the firm might have? Which models do we use where? Where do we allow it to launch sub-agents? Where do we not allow it to launch sub-agents?"

Danielle adds that enterprises are not dealing with one harness but several at once: direct model access, productivity-layer harnesses, vendor tooling, and the bespoke systems a team builds per service line. The simplest harness, she notes, is a folder of markdown files.

## Cheap analysis, scarce judgment

> "Generating content is cheap, generating legal analysis is cheap. True judgment and accountability is scarce."

That imbalance is why Danielle expects the third wave of legal AI to be compliance systems that steer agents in real time. The first wave was tools for lawyers. The second, where most investment is now, is AI-native legal and compliance businesses. The third moves upstream, into designing the framework agents operate within so the output is compliant in the first place.

## The counter economics

The most surprising moment comes from a general counsel Danielle spoke with, who sent an RFP for an M&A deal asking every firm for two quotes. One to validate the first-pass diligence report her team had generated with internal AI. One to do the whole thing end to end.

> "Every one of those firms came back. It was more expensive to do the validation on the first pass that the client did internally versus doing it end to end."

Which brings both guests to the same conclusion, from different directions: the scarce input is the expert who can steer the system, and where that expert sits determines where the work should go.

## Transcript

Full transcript of the episode. Timestamps match the published video.

**[0:00]** *Cold open.* Danielle Benecke: "As an industry, we've gone from a what tool do we roll out conversation to where does our harness live conversation, and I think most organizations are wising up to that right now." Max Junestrand: "Take the entire service market, it's roughly 5% software, 95% service, and that is about to change. It will make the supply of legal services much more widely available. It will make it possible to pursue strategies in litigation that we otherwise may not have found or not have thought of."

**[0:41]** Welcome to "Enterprise Aligned AI," the podcast where we interview enterprise practitioners on the reality of agentic AI adoption. What gets deployed in regulated security-conscious enterprises? How is success measured? And how are the best leaders navigating the changes? My guests today are Danielle Benecke, who leads Baker McKenzie's Applied AI practice a firm with roughly 4,600 lawyers more than 70 offices worldwide, consistently ranked the world's number one legal brand. And Max Junestrand, CEO and co-founder of Legora, the Stockholm-born legal AI platform founded in 2023, now valued at over $5.6 billion and used by more than 1,200 legal teams worldwide.

**[1:29]** It's truly a rocket ship. Thank you both for joining the podcast. Thank you so much for having us, Aparna. It's great to be here. Pleasure to be here. I'm glad we could make this happen. This all came out of Future Law at Stanford, so shout out to Roland Vogel and the team for bringing us together. Totally. Legal is a hot vertical right now for AI, but both of you have been working on AI for many years before danielle, let's start with you. People picture a global law firm as being a buyer of technology, not necessarily building AI capabilities, and yet you've been leading the Applied AI team at Baker McKenzie.

**[2:03]** You built it up from scratch. What led to that, and what unique capabilities does your team bring ? So we've been, scaling our Applied AI practice since 2020, 2021, and it builds on a much bigger and longer AI program that dates back to the 2010s. But the real inflection point we saw back in 2020, 2021 started with the GPT, 3 API, and we started experimenting with that and bringing it into different service lines. And that, for us, was the, moment where we realized we needed a forward deployed capability. That's the capability that I founded, grew, continue to grow and lead now.

**[2:40]** And we are a group of lawyers and AI technologists that specialize in designing, calibrating, and steering AI systems that do complex legal work. The companies that we work with are large global multinationals that work across many jurisdictions and have really complex legal and regulatory landscapes. And so what we do is we design systems that can undertake certain types of legal and compliance work to support those businesses. Max, Legora also started in this field before we really had the foundation models, so there's a lot of vision that went into the company.

**[3:18]** Tell us, as a co-founder of Legora, what made you start this company? I don't think I get to take credit for those, first gnarly years pre-GPT, because I jumped on the ship in 2023, and when things really started taking off. But with the early BERT models, and in particular there was a Swedish version called SweBERT, that my co-founders were, utilizing to summarize and work with court cases. It was really hard. These models were not very intelligent. They were very hard to work with, enormous context issues. And then suddenly we get to GPT-3.5, right?

**[3:53]** And things just, cascade from there. And so we took the bet and we said, "Hey, let's, quit our jobs. Let's, let's drop out. Let's pursue this full time." And about one month later, we started working with the biggest firm in the Nordics. A month after that, we got accepted into Y Combinator, and then we were off. And since then, not only has the underlying technology continued to develop, but the way that the scaffolding, the tool calling, the harnesses, the amount of performance that we can squeeze out of these models, and the different user experiences that we can build around them, that has all flourished.

**[4:29]** It feels like we're running a marathon, but that we're still in mile one in many ways. Was the original vision to, summarize and automate legal work, but then in 2023 there was a breakout moment and you realized that it could be more than that? Well, back in 2020 it was more of a fun side project, I'd say. They were building their own text editor. They built their own semantic search layer, and all of this we just put on the shelf as soon as GPT-3.5 came. And if you think about w- where the vision for the company and the impact that we want to have in the world continues to develop, it's that, the legal service market is so people-dependent, right?

**[5:10]** Take the entire service market. It's roughly 5% software, 95% service, and that is about to change. And I think the impact that that will have in society will be very positive. It will make the supply of legal services much more, widely available. It will make it possible to pursue strategies in litigation that we otherwise may not have found or not have thought of. It will allow us to, level up from gruesome, doc review to the more interesting and maybe creative aspects of the work. And to Danielle's point, it actually allows us to solve some of the challenges that your clients and large organizations are facing, because at the end of the day, we're building this technology, we're applying the technology, and we're innovating because we wanna go solve problems.

**[5:57]** It's so interesting that you mentioned BERT, Max. So one of the first AI native services that we deployed to clients back in 2021 was we were taking the GPT-3 API and combining it with BERT and some other capabilities to do some scaled supply chain risk- monitoring for clients, right? Obviously a very primitive capability to what we're doing right now. But I think for us, that was a really early signal about the commercial model- that was possible. To your point, Max, this is fundamentally about new services that solve unmet needs. I think one of the challenges we have as an industry right now is most of the market is focused on accelerating existing services, existing ways that individual lawyers work at the task level.

**[6:44]** The real opportunity that certainly we see for our clients, and that we see for the industry at large, is around entirely net new categories of services, and that has been our focus from day one of our program. Danielle, you've been at the forefront of both, the legal challenges that AI cr- creates, as well as the types of legal problems that AI can solve. Most people think there's potential for AI to do the rote work of, the legal profession. But you're saying there's new types of problems and new types of work that maybe we haven't seen.

**[7:17]** Can you give us a few examples? Yes. And somewhat counterintuitively, I think the true opportunity is around those more complex categories of work. A lot of the market sees the opportunity more around accelerating lower level work. I think the real opportunity is around complex work that humans were simply not able to do without this toolbox that we have now. The other big thing that we're seeing in our practice is a lot of the work we're doing with clients stems from them rolling out AI across their businesses. In the very early innings of the shift, we had a number of clients come to us and say, "Marketing and comms have been the first to roll out AI across the business."

**[7:59]** They're now creating massive, mountains of legal work that need to be done because in the old days, they partnered with the advertising agencies, and they would produce tens of pieces of content that our legal and compliance r- teams would review. They're now cutting deals with the frontier labs, and they're producing tens of thousands of content that is highly customized down the long tail of countries that we operate across. So one of the other early service lines that we got into is AI-enabled triage for global marketing and comms content for a lot of our global multinational clients.

**[8:33]** And we would be analyzing this content across Product regulation across IP risk, across ambush marketing, and all sorts of other legal and regulatory verticals. But we said to the clients very early on, "This is not a scalable way to solve this problem. What we need to do is move upstream with you and start to design compliance systems that can help steer your marketing and comms agents and other agents operating across the business." So there is compliant possible in the first instance, right? One of the big challenges we have right now is generating content is cheap, generating legal analysis is cheap.

**[9:11]** True judgment and accountability is scarce, right? And so we don't wanna get in the situation where we're in this AI arms race, where we're just generating more and more content that the legal and compliance teams need to deal with downstream. We want to improve the quality of action across the company, and ultimately, I think the third order effect or the third wave of AI in our industry is going to be all about runtime compliance steering for companies that, have agents operating across their business. The first wave was obviously tools for lawyers.

**[9:44]** The second wave is now AI native legal and compliance businesses. That's where a lot of the investment activity is now. I think the third wave is going to be about designing compliance systems that can help steer agents and business behavior in real time. That's fascinating. Designing, compliance agents that are, steering, output in real time, sounds like it requires a collaboration between legal practitioners who, understand regulation, and all the nuances, and AI engineers. How do you make that collaboration happen? We are seeing the early sparks of this right now in our industry, and to take that example earlier of the marketing comms compliance, where the original service was the downstream triage- the next iteration of that service is around designing a framework for the agents to operate within so that the content is as compliant as possible in the first place.

**[10:41]** And that starts to bring the walls down between legal risk compliance silos in the business as they historically have operated into, a much more collaborative en- environment. We're already seeing that happen across a lot of big global enterprises. Does Legora, which is a technology provider for the legal space, versus Baker McKenzie, which is a law firm do you, work very closely with your customers to start to invent the new aspects of the application? I think Legora is one of the only technology companies where we have more lawyers than software engineers.

**[11:16]** It's very, very rare, but we do. And we've actually, I don't know who, who invented the term first, but we've certainly adopted it, and it's the legal engineer. And the legal engineer is a very tech-savvy lawyer who both understands the problem that we're solving in terms of the use case, but also is adept enough to actually help drive change management and the transformation of how legal teams operate. Because if you just take this technology and you go, just as Danielle said, "How can we improve, the way that we already do things?"

**[11:50]** you'll get one type of result. But if you Say, well, now we have this technology, how should we reinvent the process all to begin with? Then you can get something much more, a- aligned with the capabilities that we now have. And I think in terms of new things coming, I love where this is going because effectively, all of the agents operating within an enterprise are going to have to, in real time, be able to query, is this plan or is this thing that I'm now going to do compliant with our policies or compliant with regulation?

**[12:21]** And the regulatory, body that you actually need to take into, account is enormous. It's too big for one single person to actually keep in their brain at all times. There's just too much context. And as we start to build these layers for agents that can operate more autonomously, then I think that's certainly a kind of cool path to keep going down. There are all these new roles, forward deployed engineers, which have been around for some time, and now legal engineers who are lawyers and also engineers. But I wanna ask you, Max, about differentiation and moat.

**[12:53]** This is a really important question, especially in the legal profession, because there are many different types of players. There's all these incumbent companies which have distribution like LexisNexis and Thomson Reuters. They can obviously pull in AI capabilities into their existing services. And then there's a range of startups. Legal is a very hot area for, funding startups- all going after different parts of the market, and certainly law firms are a big part of that. And then there's the question of can foundation models themselves start to, increase what they offer by, post-training, for example For Legora, what are the true sources of moat?

**[13:31]** Well, the space is developing at an exponential pace. And I like to say that velocity is a moat because you have to be at the forefront, and you have to be at the frontier at all times. If you look at what our product was back in 2023, it would have looked like a total joke today. If you brought out the product from 2024, it would have looked like, a really poor executed legal AI platform, and maybe even towards the mid of 2025. And with the things that we've just come out with, we are, clearly moving ahead of what I'd say the most of the pack is doing.

**[14:10]** And as you said, everybody's thinking about how can we win our, slice of the cake here, and we're still in mile one, right? Technology and product tends to compound. So if you make a lot of really, correct strategic decisions about where should we train models, where should we not train models, where should we build, where should we partner, where should we buy? How do we build a user experience at the end of the day, and a product that delivers, more value than anything else can do on the market? And that's what Legora is obsessed about.

**[14:42]** There's, I'd say, different strategies at play. Yes, you have some of the big incumbents who are saying, "Oh, we're gonna, leverage this piece of the data and our big distribution muscles, and that will be our way of winning." And, fine, that's one strategy, and we see what's happening with those, public stocks. Um, you know, they're, they're maybe not doing, super hot right now. And then on the other hand, you have private companies like Legora, who is in, enormous demand, , in terms of our growth rate and where we're going. And I think that the old sort of SaaS world is also really getting disrupted by these AI native businesses.

**[15:18]** If you look at the growth rate of some of these new players entering on the market, it's just, it's throwing all of the old metrics and the even venture capital playbook, out of the window. Because the net revenue retention, the DAU over MAU, growth in revenue, this is just, unheard of. We get to experience that in legal, but the same thing is happening in coding, right? You've got Cursor on one end who just exited to, SpaceX and, you have, all of these different, verticals that are playing out, and we're certainly not at the finish line, for the legal one.

**[15:50]** Aside from velocity, what are the other moats? Well, I mean, the- there's the degree of em- embeddedness with a customer, when you work with a firm like Baker McKenzie, and you roll out to thousands of practitioners, and they start to adopt Legora in terms of how they do work and deliver to their clients, that becomes one type of moat. When you,, build products like our portal, that actually puts both the clients and the law firm- In one single software experience, that becomes a way of driving both, new types of delivery models, but which of course are quite sticky once you get them right.

**[16:25]** The way that you integrate with the rest of the ecosystem and that you build functionality that frankly is, quite far away from what a foundation model would naturally go after. A lot of the software that we're building now has nothing to do with LLMs whatsoever. , The amount of audit logs, permission management, trails, scheme. The amount of just things you need to solve in order to deploy within an organization like Baker or a big enterprise for that matter, in the legal, department specifically, are quite extensive towards what just the, horizontal models are doing.

**[17:01]** It's maybe not so clear at the surface, but the minute you step your foot in it, you realize that, oh, this pool was really shallow. If you wanna upload 1,000 documents from a VDR or 100,000 documents from a e-discovery exercise, and it just like blatantly fails in some of these frontier, models, like in the way that those are packaged in their, very general harness. And you go to Legora and it's just like night and day. We're not positioned as we wanna be the legal AI assistant for everyone. We're a agentic, infrastructure layer that works with some of the biggest enterprise and legal firms in the world.

**[17:37]** I think we're naturally going after quite different segments of the market. We've all worked with very large enterprises. There's a huge compliance burden would you say that Legora is a legal specific harness? We certainly have a harness that's tuned to legal, but there's many layers in the product. , We articulated this quite well, I think, with our agentic operating system that we released a few weeks back. It starts with the foundation models, and then we put those foun- foundation models into work within our legal harness, and then we give that harness access to legal specific tools, legal specific context, legal specific surface areas.

**[18:11]** Because not everything should be served in a chat, right? Every use case, the UX is not, a chat is not optimized for everything. And then all of that has to be wrapped in, a secure and enterprise grade system, and it has to be combined with the legal engineers. We should not forget the human aspect of this. Even if you have a tool like, Claude, implementing that within an organization to drive the outcomes you want is quite extensive. There's a lot of DIYing you have to do if you take a general harness.

**[18:45]** Probably, no organizations in the world can take that to the extent that we've taken Legora. And that's because we're 700 people who wake up every day and obsess about our particular problem. Whereas in these, very general tools, there's maybe one PM and a few engineers. But what is the harness? What is different about your harness than, let's say, Claude Code? Claude Code is optimized for coding. I-i-it goes back to how do you let the model read the documents, make sense of documents, make sense of context? How does it tie together?

**[19:20]** What's authoritative between that case over there and some internal precedent that the firm might have? How do you actually weigh those two together? Which models do we use where? Where do we allow it to launch sub-agents? Where do we not allow it to launch sub-agents? How do we instruct the agent to actually plan? We recently, introduced a new agent, and that is tailored to create a very detailed plan for the user before it sets off to do the work. Cursor does the same, right? When you're coding with it, presents a plan before you actually a-allow it to go out and execute.

**[19:54]** But then the way that it plans is, of course, a function of the tools at its disposal and, subject to its own capabilities. I think you can go very, very deep, and I could probably talk for an hour about, where we wanna take it Just to jump in on harnesses, it's important to distinguish between inner harnesses and outer harnesses. One of the things we see in the narrative in the market right now is harnesses becoming a monolithic concept, when in reality we're actually dealing with multiple harnesses if you're a big enterprise business.

**[20:26]** So we, for example, have direct, access to a range of models through our cloud AI environment. We work with open AI models, we work with Anthropic capabilities, we have access to Google capabilities, et cetera. We then have access to productivity layer harnesses and Anthropic Cowork is a great example of that. You see similar capability filtering up through Microsoft environment as well, and then you get into legal specific tooling. Most large scale law firms and other actors in the knowledge industry have access to multiple tools with different types of harnesses. As an organization, you also are starting to create your own harnesses.

**[21:08]** Every service line, every bespoke system that we design and deploy for our clients, that has its own unique harness. The most simple harness is probably markdown, a folder of markdown files, right? A big piece of work, that my team and I do is working with chief legal officers, chief compliance officers, general counsel, other leaders on the client side, and help them design their stack, their strategy around this technology. They're starting to think about the strategic dependency of having their harness live inside another platform, they're starting to think about, what harnesses should they own rather than rent.

**[21:45]** One way that we think about it is there's a scale from generalizable to custom. There's a scale from extremely strategic to not so strategic. And I think for the stuff that is super generalized, not particularly strategic or important to your organization, totally fine to rely on a harness that lives outside of your environment. But for the very strategic stuff that ultimately turns on your expert steering or internal playbooks and internal know-how, you need to be very careful about where that harness lives. As an industry, we've gone from a what tool do we roll out conversation to where does our harness live conversation, and I think most organizations are wising up to that right now.

**[22:26]** Does Baker McKenzie think about, AI adoption as a top-down mandate or a bottom-up mandate? You're using many different tools, many different models. Is that , by choice from, lawyers within the firm, or has it been something that's top-down? We're a services business, so everything starts with what our clients need. And in this era, it really comes down to three things. Our clients need strategic clarity, decision velocity, and risk mitigation if you look at every piece of legal or compliance work, it's ultimately those three things. As an orchestrator of capability that serves that ultimate need, what do we need?

**[23:03]** That really drives every question about the capabilities that we plug in, whether that is our baseline cloud AI capability and multi model access, whether that is bringing in best-in-class vendors, whether that is Legora or other players in the ecosystem. The market is moving so quickly, we need to be very agile, not bet the farm on any one thing, but just c constantly thinking about what is ultimately going to serve the client need. Many players in the industry still think of this as a primarily a technology shift. Technology is obviously a big part of it, but the commercial model transformation part is arguably much more profound than the technology shift.

**[23:44]** When I look to what we build and sell to our clients, the technology bit in many ways is the easy bit. The hard bit is how do you sustain that capability in a way that is commercially viable for the business and for the client, and that also meets the governance needs of sophisticated global companies, when my team is, building out this set of primitives that ultimately drive all of our service lines, the commercial and the governance pieces are equally, if not more important than the technology pieces that we bring in.

**[24:19]** I see. So depending on client needs, you'll bring in the relevant technology. But the technology shift is actually leading to a commercial model shift. Tell us more about the commercial model shift so historically, knowledge work has been about, execution, human workers executing on certain tasks. There's no secret we're five years into the shift at this moment. More and more knowledge work is about designing systems that execute rather than the downstream execution. Many lawyers and compliance professionals still are thinking of their role as shifting downstream. A lot of firms and other players initially thought, "Well, we'll have the AI generate output, and then we'll validate downstream."

**[24:58]** The most important expression of expertise at this point in time is actually upstream in designing the system. The better the design, the better the output is going to be, the less downstream validation you're going to have to do. That ultimately changes what we're selling completely, and that ultimately creates an en- entirely new model, commercial model for delivery of work. So for example, my team, we operate off an entirely fixed fee model. Every engagement, we will design and price that around what the client needs. What kind of risk are they trying to mitigate?

**[25:31]** What kind of commercial opportunity are they trying to unlock? What is the ROI they're looking for? And then we design the service around that. Very different, pricing model to the way that historically legal work has been delivered. So it's not about the billable hour and the expertise. The expertise is, like you said, better applied in the design of the solution, and so you scope out based on a problem what the solution is, and then it's a fixed fee, which is more outcomes based. Exactly. And working within a large incumbent law firm, moving the ship takes time, right?

**[26:06]** My team operates in this way. That does not mean that every team across my firm is operating off of the same billing and pricing model because it takes time for a wider industry to shift. But for the AI native services that we sell, that is the way that we think the commercial model, is best shaped. If we look at the legal industry and the transformation that it's going through many of the new YC startups are totally AI native, where they're providing a legal outcome, and, it is a black box in some sense, whether it's done by AI or it's done by lawyers or who it's done by.

**[26:39]** Do, do you also, price based on outcomes? First off, we're not an AI native law firm, we're a software company. And I wouldn't be, I wouldn't wanna compete with these guys. They're very competitive. I think that the AI native l- law firm model, is tricky because you're actually competing with the full-time software builders, and you're competing with the full-time experts who have built their entire, expertise, customer base, processes around delivering legal services. So I think it's a hard market. How do we price? Well, we've priced primarily on a seat and enterprise basis, but the move to introducing more and more agentic functionality, has actually made us introduce consumption-based pricing.

**[27:26]** Most AI tools since the, beginning of Cursor and Lovable and Clay and also the big frontier labs have been on a consumption basis because that's the only thing that aligns the value delivered with, the cost basis. We don't have software margins. These AI businesses, are fundamentally built in a different way, and this is the only way to make sure that if one user is leveraging, hundreds of dollars worth of tokens or even thousands of dollars worth of tokens, that has to be priced differently to a user who's doing much more, foundational and basic things.

**[28:01]** We spoke about moats earlier, and you mentioned speed legora has grown exceedingly fast, you mentioned that there are certain important, decisions, and pricing being one of them, but there are other important decisions. So tell us what helped you grow? Thank you. We have grown very fast. We formed the company, in the late spring of 2023. We were accepted into Y Combinator in '24, and then from 1 million to 100 million, I do believe that we, beat the record for the fastest enterprise company, to go from 1 to 100 million in less than 18 months.

**[28:31]** We were, like, very on, on a knife's edge to beating Wiz. But I think we didn't with about a week or two to go. And I think early, we made some tactical decisions that we were going to focus all of our attention on the application layer, and that we were going to build with our customers. So I'm a software engineer. The entire founding team were software engineers. We didn't hire a lawyer at Legora until six months into the journey, and he also happened to be a software engineer and a lawyer, so, he's a little special.

**[28:58]** We have three cultural values at Legora, and the last one is grow together because it's so clear to us that we only win if our clients win, and we need them emotionally and commercially invested in how this technology will be brought to, the field. And so every new piece of functionality that we build, we co-design and build with a group of design partners. We run a really tight customer advisory board. We iterate a lot, and we have a saying that when some zig, we wanna zag, and that's because you can't just look at what the competition or what others are doing.

**[29:43]** I think there's a lot of copying And if you're swimming a race and you're looking sideways all the time, you're gonna be slower. So you need to look down. You need to do your own thing, and you just need to move really fast. And one of those decisions about being one of the first major players to come out with consumption-based pricing now is also because it allows us to continue being the product and innovator in the space. What's actually quite good about it too is maybe in comparison to some of the, general and big tools, we can work with a much, finer, group of models.

**[30:16]** We don't have to put Opus 4.8 into everything we can use open source models. We can use, older generational models because once something is good enough at a specific task, you don't actually have to push the frontier models anymore. So we really wanna align our commercial interests with the commercial interests of our clients, and then be very transparent about how it's going to continue developing forward. So you can potentially be more efficient. Token usage is right now, quite a premium, especially for the larger- Yeah. We don't want people token maxing in Legora, right?

**[30:46]** We want them to, get the results they need. And as a company, we started in, in Sweden because that's where we happen to be born. But very, very quickly we've become a global, company. Maybe not as global as Baker, yet. Not, not yet. But we're getting there. We have something like 15 offices up and running all the way from San Francisco to Japan and in between. It's also because, different markets needs different things. The way that you work in Singapore is different from Germany, which is different from the US, and we've really, recognized the need to be local, in these markets.

**[31:19]** So what led to Baker McKenzie choosing Legora as one of your, partners? We can talk about that in a sec, but I actually wanted to jump on, the tokens point because I think it's a really important one when it comes to the commercial model. And to Max's point, we are well out of the AI subsidy era, right? We're past the token maxing phase where we all had our Apple Minis running all weekend on- all sorts of, even just our personal projects. We have started to see that, come up as an important factor in pricing services.

**[31:50]** So to take one example, we do a lot of AI-driven global compliance audits for clients. Those can generate billions of tokens of analysis once you're working across multiple verticals across many countries. That's becoming a meaningful part of the cost equation for delivering this work. So that's just another reason why the pricing model needs to change. Ironically, I think, Max's announcement about, the consumption-based pricing, which is, very expected, it takes us out of the earlier narrative about legal, AI because I think a lot of the clients thought, "Oh, we're gonna get out of hourly-based pricing, which is super unpredictable."

**[32:26]** Yeah. Now the vendors are coming up with a even more unpredictable pricing model, which is tokens. For now, that's why we are very focused on fixed pricing. We do a lot of work to model out what we think those cost inputs are going to be before we scope the work and deliver it to clients. I think this is a kind of natural part of the shift that we're going through, and most enterprises are going to have to figure it out. The other thing that's a little bit counterintuitive is when you're looking at the relative cost of different models Different models have different token efficiency.

**[32:59]** And so sometimes you might be looking at a more expensive model, and the instinct is we'll go with the cheaper model that's most performant, but actually the expensive model may be much more cost efficient. It's like a great shampoo. Exactly. Sometimes it's better to get the slightly more expensive shampoo because you end up using less. You use a little less of it. I agree that introducing consumption-based pricing introduces another, metric of uncertainty. However, once you have the data, it's actually very predictable. If you're running a big, say, tabular review on, a thousand documents x 100, queries, we know what that's gonna be.

**[33:33]** And similarly, if you kick off a big agent workflow- we can predict with pretty good certainty what that's gonna be. Now, some firms like Baker have a really strong pricing, team, and you have a lot of data in the firm to know how to price stuff, but not every firm has that. And we already have, certain litigation tools that get passed, right? Like The cost of certain discovery tools gets passed down to the end invoice, I think some firms might experiment with, doing the same for their AI usage. I think that's right.

**[34:03]** Then going back to your question about how do we orchestrate different vendors in our stack, it's really about what do we think is best in class capability to bring in. As I mentioned before, we're not gonna build something that we think is generalized capability that is best served up by another player. Openly, we have our cloud AI layer, which plugs into all sorts of capability, whether that is Google, Anthropic, OpenAI, and other models. Like many in legal, we are primarily an Azure shop, but we do have other cloud AI environments.

**[34:33]** I am a technology lawyer by background, so a lot of the work that I do actually is with, the legal teams of Frontier and other tech driven firms. Sometimes those clients wanna use their own models for obvious reasons. We can meet those clients there. And then as we get up the stack into different capabilities, we use Legora for certain things. Certainly in the transactional realm, we're seeing some really, useful capability to plug in. We use other specific tools, whether that is Relativity Air for very domain specific things. My team will often layer A lot of capabilities within the same service line.

**[35:07]** To give you one example, we support a lot of cyber incident, global cyber incident response. So one of the services that my team delivers is rapid same-day, analysis of reporting obligations for very large cyber incidents. Historically, it's taken days or weeks to do that kind of analysis, triangulating across different regimes, especially if a client is dealing with an incident that affects millions of consumers across multiple countries. So we have our own internal harness, which is built up of all sorts of logic and playbooks, that drive that service. We use foundation models directly from some of the labs to power that service, but equally as an input if we wanna run a forensic on the incident, we might use a tool like a Relativity, and then we will take that forensic report, and we will feed it into our system, and we will analyze that report against our playbooks.

**[36:02]** We see the future of these kinds of AI-driven services as just increasingly complex stacks of best-in-class capabilities orchestrated, with the right commercial model around it, the right governance around it to give the clients what they need, which is strategic clarity, decision velocity, and risk mitigation. What you're delivering there is, fundamentally a new value function. Yes. Because if you have a cyber incident, you are willing to pay more if you can get the results tomorrow than if you have to wait three weeks. The technology savviness of the, firms is increasingly becoming a competitive advantage, I appreciate that, Max.

**[36:44]** The other thing is the service completely changes. So in the old days, the service would be you do the analysis once. Today, you can't do that because most of the cyber incident reporting regimes range between six hours deadline to 48 hours deadline, and you are getting new facts, new assumptions. You know, the boom happens, and most organizations get the information in piecemeal. What we're able to do is just constantly model the reporting obligations as the incident develops. So we might run the analysis many times over the course of 48 hours, so the client is getting the best real time analysis of what they should do.

**[37:24]** What I'm hearing from you, Danielle, is that, there's a range of problems that were difficult to solve, previously. Now that you have this, riches of tools, including the foundation models themselves, you're able to build solutions that didn't exist before, you're pricing for those solutions based on outcomes, which bakes in the consumption piece. What you're surfacing here is a tension with AI that, all of us feel, which is that because token consumption is such a big part of the cost, you have to pass through some amount of consumption on the other hand, you're providing new kinds of capabilities and you want to price for that.

**[37:58]** Once you see some of the hourly rates, you realize that the tokens, in a lot of legal work delivered, the tokens are a very small piece compared to what some of the hourly rates are. That's a really interesting point as well because we operate across 45 markets that have very different historical hourly rates and very different economics around legal work. I was in India recently for the AI summit connecting with a number of our clients out there interestingly when you talk to the chief legal officers of those companies, they will say it's actually more efficient for us to get our local expert to weigh in on something than to have a expensive US commercial reasoning model mull over it you have to be really thoughtful, particularly in this AI scarcity era, about the kinds of problems that are ripe for this approach.

**[38:47]** I was talking to, a chief legal officer of a technology company recently, and he framed the calculus around whether he will send work outside the organization or keep it in around three things. He said The primary question is, do I have the right expert to steer the AI system to the right conclusion? If I am not confident that I have the right expert inside of my in-house team, I am likely to get a lot of slop and a lot of token spend that is just not efficient versus sending it to an expert that is proficient at steering the AI system to get the work done.

**[39:25]** Secondly, even if they think they have the right expert, do they really want to hold on to the risk of that work? The accountability question, do they want to send that risk outside because it's a kind of decision point that they need, somebody to be accountable for. And thirdly, what is a overall kind of cost efficient way of delivering this work, taking those two things into account? And I thought that was a really interesting way of framing it. The other thing I've seen recently, this is another general counselor who said that to me.

**[39:53]** She said she had sent out an RFP for an M&A deal, and she said to all of her firms, "Give me two quotes." So one quote is for you to validate the first pass diligence report that the client generated using their internal AI. Second quote they wanted was just to do the whole thing end to end. Every one of those firms came back. It was more expensive to do the validation on the first pass that the client did internally versus doing it end to end, I think it signals that where we're at right now is it's all about do you have the right expert to steer the system to get the work done efficiently because we're seeing all sorts of counter economics.

**[40:36]** Wow. My question was going to be about build versus buy. It's not about build versus buy anymore because, you're choosing the right tools, you're creating custom solutions. You're seeing a whole new set of problems that you can solve that you couldn't solve before. The question is really about where is that boundary? This is a question to both of you, because you have to decide what your roadmap is, what you're going to build, and what you're not going to build. Yep. You've mentioned that, you are working very closely with customers and, you're charting your own path.

**[41:01]** Not looking at other swim lanes. At the same time, every client is different you have to say no to things. Yeah, my main job is saying no, actually. So where is that boundary between- what you would build into Legora versus what you would build a custom harness for- I think the first thing to mention is, we are very transparent with, our clients about what we are going to build and what we are not going to build. We actually recently just got sent an Excel from one of our close partners with, "Here's all the software tools we have.

**[41:31]** Mark an X where you think, over the next three to five years you might do something in this space." So we can, plan our software stack, because, it's, clear that, Legora is growing to be something quite big. The part around, what do we decide to build and what do we not decide to build, the most important thing is that on top of Legora, you need to be able to configure and build the differentiation. So if Danielle and team wants to bring their own MCP server into Legora, they can do that today.

**[41:57]** But they probably don't want to build the UI- UX and routing and orchestration of a Microsoft Word add-in, 'cause that's, really annoying and very gnarly, and we get a lot of scale on building that. But building a custom cyber solution for a very specific type of cyber case, that's maybe something more up your alley. It's also a timing question, right? Because, our clients can get arbitrage by having certain capabilities now versus having them a year or two years from now. Because technology is becoming an increasing competitive advantage in how we pitch and how we win work, buy versus build for me is more of a how do we partner and how do we make each other as successful as possible?

**[42:41]** Anything to add to that? We do a lot with our clients on this, help them think about where to build, because even the most sophisticated technology companies in the world, their legal and risk and compliance arms, it does not make sense for them to roll their own for everything, even though when they have the frontier models inside of their ecosystem. One of the ways I think about it is on the spectrum from generalizable to customizable, where is that capability? Typically, the more generalized it is, the more that this is something that other firms, other players in the ecosystem would benefit from that kind of capability, it makes sense to leave that to the wider market and to buy that in.

**[43:20]** The more bespoke it is, the more it turns on our unique expertise, our unique playbooks, things that others do not have in the market. That's going to make us more inclined to keep that capability internal and to have our own harness, our harness around that. The other axis would be around strategic criticality. How important is this to our business versus is this a fairly commoditized capability that others can replicate in the market? So the quadrant where my team typically tends to focus is complex, high-value, strategically critical services that turn on our expertise, our playbooks, things that nobody else in the market has access to.

**[44:01]** Generalized capabilities we would typically buy in, but everybody's stack is going to be different. Ultimately, I think that's where most enterprises will land. Great. That makes sense, and I think you're both aligned on that. One last question, There's so many companies offering all kinds of AI services, of course, legal services these days when I talk to CXOs, they're like, "Well, I don't know who to go with because there's so many vendors. I can't tell the difference." You've obviously been able to cross that hurdle. What do you do? What's your secret sauce for winning over clients?

**[44:32]** Except for Dude Law, of course. It all started at a very, from a very honest place, we wanna go build delightful software with the best lawyers in the world, and that's the pitch, and that is continuing now as we scale. Of course, you cross these critical thresholds, you get to one million revenue, you get to 10 million revenue, you get to 100 million revenue, and then the next one, and that builds your brand, and that builds, interest. But ultimately, buyers are comparing many solutions. And so the only way that we actually continue to stay relevant is by being the absolute best product on the market, when we go into a competitive pilot, we have to out-compete other providers to win the trust our clients.

**[45:19]** And that comes down to product, that comes down to the partnership, that comes down to the roadmap. And as Danielle said, you have to build, trust, and you have to think about how this is a critical piece of infrastructure. And so the company that you decide to work with, you need to make sure that they have their heart in the right place. I think time and time again, you have to prove that and build up a reputation. Given the trust aspect, it's very important to actually check the work of the AI, make sure that it's correct, right?

**[45:44]** Yeah. You've mentioned before that, having the right expert, anywhere in the world. There's a very human aspect to that, at least for Baker McKenzie. Would you say that that's part of the edge? Absolutely. I think most of the market is very much focused on the downstream validation. For us, it is equally, if not more important, to have the experts designing the system in the first place. That is where you get the more performant systems when you have the right experts designing the logic, designing the playbooks. Because many of the playbooks that power our services, it's not pure black letter law.

**[46:18]** It's a pragmatic, risk-calibrated approach to the law. When we're advising our clients, we're not just saying what the law is, we're advising the clients what to do in the context of not just the legal landscape, but the market, their product roadmap, the political, geopolitical landscape and society. And those are two very, very different products. Validation is important, but upstream and downstream validation need to happen. I think also AI actually gives the first opportunity to, scale this across bigger firms too. 'Cause if you have an amazing partner in one of the offices write a skill for how an agent should do a specific piece of task, that can now be scaled across the entire firm.

**[47:00]** So actually expertise and precedent and knowledge and knowhow, can in a way get scaled across the business in a much more pragmatic way than ever before. This was a very energizing and rich discussion. Anything else you'd like to say? Anything about the future of law? Internally at Legora, we like to say that it is the best time in history to be a lawyer because never before has this many hardworking, devoted, and talented people been focused on building great software to make it more fun, energizing, and deliver better results for the clients.

**[47:35]** So I'll leave all the listeners with that. I would agree with that. I think it is the most consequential moment to be a lawyer or a compliance professional. I would encourage the industry to focus more on what are the services that don't exist now that could be created with this capability. We have spent the last four or five years as an industry focusing on accelerating existing services. There is value in that, but I think far and above the bigger value that we can deliver to the clients, to the companies, is around unlocking some of those existential risks that keep the chief compliance officers up at night, that keep the CLOs up at night, which are things we haven't solved for quite yet.

**[48:19]** Thank you so much. Thank you. Thank you. It's been fun.