#61 If It's Scary, It's Good How TravelPerk's GC is Reimagining Legal Services

Hello everyone and welcome to Pearls on Gloves Off. I am your host, Mary o', Carroll, Chief Operating Officer at Goodwin. Today I am delighted to be sitting down with my friend and true innovator Andy Cook, the general counsel of TravelPerk. If you're not aware, Travel Perk is a hyper growth global tech unicorn SaaS company that provides a business travel platform as a pioneer in the future of business travel. It offers an all in one platform that allows travelers to have the freedom they want while providing companies the control they need. At TravelPerk, Andrew heads the risk management function and his team includes personnel in London, Barcelona and Berlin. Before being General Counsel at TravelPerk, Andy Built Award winning teams for other companies involved in creating connections for people, including at Fnatic, the leading performance brand in esports and a pioneer in the pro gaming space. Before that, Andy spent eight years as GC and Head of Strategy for Flash Entertainment in Abu Dhabi. He and his teams have received countless awards over the years for Innovation, Department of the Year, GC of the Year, and so on. I think you will see why in our conversation today. For that same reason, Travel Perk is the first organization to have two employees now be guests on this podcast. I have met so many members of this team and there is truly a high level of talent, but more importantly a real mindset focused on the future here. In today's conversation we cover so many things about AI, the future of law, the changing client landscape, the innovator's dilemma, how to train associates for the Future, and what TravelPerk is actually doing with AI today and so much more. I always enjoy my conversations with Andy and learn so much. I hope you enjoy today's conversation as well.

Goodwin is wherever AI is, which means we'll be at HumanX, the number one AI conference for business leaders. Join us on March 9th in Las Vegas for exclusive panels covering foundational, legal and commercial topics pivotal to those navigating AI, including what to consider before making your next business decision. Visit goodwinlaw.com humanx for more details.

Hey Andy, how are you? It's always good to see you Mary.

It's an absolute pleasure, honor and a privilege. I feel like I should have a tuxedo on or something to be meeting legal ops royalty. A real pleasure to be here and lovely to see you again.

Yeah, it's great to have you. I'm actually surprised I haven't had you on earlier and I've wanted to, but you know I had one of your colleagues on and I was like when's the right time to bring, you know, the second person from the same company on. But every time I think about who we should have on the podcast and who's actually doing really interesting things, I come back to you. So I'm glad you're here.

Me too. Me too.

Should we start? You and I have known each other for a few years now, and I think I was trying to think back the first time we met because I had attended this conference in London called Crafty Fest and you were on the stage. We didn't know each other then. You were on the stage giving the keynote address and I don't know what year that was that we were talking. You were talking about AI. It had to be pretty early in the AI days.

We were pretty early in terms of using AI, mainly because we built the pathway to using chatbots. Right. So we were using chat, sort of what you might call dumb chatbots, decision tree based chatbots to deploy answers at scale to the TravelPerk team. And so it wasn't so much of a big lift for us to move into, okay, let's power those chatbots through AI. And so, without really intending to be necessarily, quote, unquote, innovative in the AI space, we found ourselves all of a sudden at the forefront of that. We were the first customer of a number of startups working in the AI space at that time. So the purpose of that KeyNote back in 2023 was essentially to examine the question of customer centricity and some of the things that I was seeing on the demand side at that time. So very often, I think when we're talking about the evolution of legal, we look very much about the supply side, as in what it is that we're going to do to change our solution. I was really much more focused there on the demand side. And there was three particular headwinds that I was talking about at that time, and those headwinds were okay. The concept of the unreasonable client. People always talking about how unreasonable clients were without seeing the massive amount of change that had happened in the service delivery space everywhere other than legal. Again, you saw the app revolution, you saw the changes that the BlackBerry and the iPhone and the productification of everything through apps and location based services. This incredible change in the convenience, immediacy and visibility of services for consumers, and then how legal tooling and service delivery had just not kept pace with that. The delta between those two things people were experiencing as unreasonableness, but actually it's just the rushing in of the convenience economy. So that part of it then how people were consuming knowledge which we could see was starting to change. And again, lots of people listening to this and I'm sure yourself have experienced a set of circumstances where, you know, someone asks you a question and then five minutes later they're in your inbox saying, oh, I Googled it and this is what I'm going to do. Like, I found the answer. So we've stopped being people that retain information, We've started being people that index information, which points us towards this idea of on demand as being this kind of key theme in terms of service delivery. And then finally, of course, and this is something which is only accelerated since I did that keynote, the diminishment of expertise, which is, again, this idea. For anybody who's listening to this in the uk, you might remember a politician called Michael Gove saying, we've had quite enough of experts. But again, now the almost active challenge of experts. This is a little bit of Paul Graham's founder mode, right? This is a little bit of Marc Andreessen's techno optimist manifesto. These are all things which essentially go challenge expertise wherever you find it. And these kind of monolithic institutions are the first place you should be going to really, actually aggressively address what might be a set on growth. If someone's saying no to you and you feel that the answer yes creates growth, then you should go with the yes. So those were all things which certainly we could see. And I talked about how we responded to them a little bit in Travel Book. But again, my only conclusion I could make at that stage, I had a sense of where it was going with AI, but I didn't necessarily see the full potential of it at that stage. But I'll admit, my big takeaway there was it turns all those headwinds into head gales, right? It was just going to be an amplifying effect on all three of those things. The convenience, access to knowledge and then this challenging of expertise. And I think that's played out. I think other things have also subsequently played out in AI, and I'm sure you've seen those effects yourself. I'd love to hear about that. But that was where I was at, at that stage of things.

I mean, that's how you and I met, because you gave that talk in 23 where AI was brand new and people were like, what is this thing? And here's Andy on the stage talking about, hey, not only have we figured out how to use it in legal, but we're. We've got all these use cases and even now, you know, we're in 2025. I find it rare that you talk to lawyers that are actually using it for the practice of law and have practical applications. I mean, certainly everyone's using it as an efficiency tool, but when we're really thinking about like, who is able to use it to change how they're practicing law. You were talking about that in 23 and that stood out to me. And so I ran up to you afterwards and I was like, who are you? Why haven't we met yet? We need to be talking. And you and I have stayed in touch ever since then. What was it? I mean, you mentioned kind of the chatbots and the early innovation that you were doing in your team, but what was it about technology or the culture that either you walked into or you created at Travel Perk that made it so ripe and ready for you to sort of like dive into what was very new at that time and then obviously to continue to build upon?

Oh, great question. Well, first of all, and I always try to acknowledge this, particularly when I'm in places where there might be people who are finding it a little bit harder to put themselves in a place of innovation. It's an immense privilege to work in a business which is growing at the pace that Travel Quirk is growing and has the open mindedness with regards to change that Travel Quirk has. As a gc, when you're coming into a team, if you see yourself as a team builder or a culture builder, to have effectively a pretty and say, open checkbook, clearly there are limits, but more like a mentally open checkbook. Okay, you want to do it that way that doesn't look like the way that everyone else does it, but that's okay. We trust you and we'll back you. That's an incredible thing to have. Right. You know, we're planting our seed in fertile soil, which I appreciate that not everybody has. And particularly now where again, thinking about that wider messaging, I'm always very conscious to think about influences outside of legal and legal. I'm not talking about just in the company, but I'm talking about just more broadly in society, but this labor arbitrage which is now happening as a result of AI. Okay, certainly I've had conversations with general counsels where they say their CFOs, controllers are looking at GNA and now saying, hey, actually we see you're spending money on these tools. When are we going to start firing people? Which is a really brutal trade off and a very weird way to create a culture of innovation. Because in the subsequent year in 2024, when I did the keynote at Crafty first, I was talking about how I reject efficiency as a driver for the adoption of AI as an industry. We've been very keen to get into tactical applications very quickly, and that's fine. Again, we want to be innovative and we want to get on that train and we want to show that we're ready to embrace these tools. One thing that we saw was a very big shift towards Copilot, which is, again, for my money, not really a product that I would want to invest in. I don't think it really creates great gains. Microsoft's a very wonderful company. It will get there with something that's wonderful for lawyers, I have no doubt. But right now it's not really adding that. And there was a very clear sense that people went into that place of safety, like what's inside of my institutional procurement net, like what are the things that I can just reach for to say that I'm being innovative? True innovation starts, at least from my perspective, from engaging with customers. And again, thinking about the demand side. What do customers not want, not what's the version of legal that we want to deliver because it meets our needs, right? Or indeed it aligns with our incentives. That is something that we very much have to shift around as we move into this space where actually we are being disrupted. We're leaving ourselves right for disruption. Again, going back to the idea of this delta between how services are experienced in every other part of our lives and how we execute against those things and the proxies for value that we've used to hide that gap. So again, typically in premium law, that can be, I'm available to you 24 hours a day and so is my team. And if you need that doing tomorrow, we'll drop everything and we just do it and we just throw bodies at it and so on. Those are the things that AI really gets into innovating against. Those are the things where even if you are very good at turning things around very quickly, your team is agile in the broad sense of the word. Right. You can just shift resources super quickly into doing things and maybe that can be tech enabled. It doesn't be on demand. It doesn't be on demand. So there's a very big bet placed that people are going to be prepared to pay the additional cost that comes from using premium law versus getting the answer today. If we start measuring this sort of thing in, I can have the answer now, and it might be somewhere between, depending on which model you're using, blah, blah, we can come back to the models in a minute, it might be somewhere between 50 and 100% accurate. Right. And again, remembering that not everything needs to be bespoked and not everything is better farm and not everything needs that degree of accuracy in order to move forward. Most things are two way doors. Even when you're consulting legal rights, how much are people prepared to really pay to get those extra final certainty on accuracy? And I think within in house, teams that haven't built on the assumption they need to replicate a private practice workflow in house, which some teams definitely do. Teams that are able to think in innovative terms are probably already there because again, the incentives drive us towards solutions that work with a very limited degree of visibility on say, the full extent, the fact or all of the different angles, the secondary order, things maybe we could have thought about may potentially impact are really now starting to kick in from it. And again, I'm sure that is also true in your world and as an innovator within law firms, I'm sure you are seeing these things all the time and challenging the thinking of your colleagues. And even from an operational standpoint, thinking about how you force project in a completely different way. Like what are you seeing in those kind of spaces? How's it affecting your workflow?

Yeah, well, first let me react to you saying what a gift it is to have an innovative culture to walk into. And I feel very privileged to have done the same thing. I mean, I never thought there'd be a law firm that had that sort of like open come in, bring your crazy ideas and have at it. And that's really kind of how I feel I'm operating so far. You know, the other thread I wanted to pull on a little bit is this customer or client centricity, which is so critical and there's a lot of different directions we can go on that. But you know, you're actually in an interesting seat because you are both a client as well as you have your own clients as a general counsel. Right. Are you seeing or have you felt changes in the role of your client or how they're operating, such that it has driven change in the way your legal department services them?

Let me ask that in a slightly different way, which is that we had a lot of skepticism with regards to even within the supportive culture. We had a lot of skepticism from the front line when we started to move into an almost entirely automated and now automation plus AI LED workflow. But once customers started to feel the benefits of that, and particularly thinking about benefits expressed in terms of their metrics. Right. So again, classic ironclad metric, time to close. Right? But again, when we start to communicate, speed advantage. So initially, speed to close on the main agreement, but also then thinking about that as only being one part of the contracting process. So you start with, if we think about the entire funnel from a customer perspective, the actual outreach starts like minus 18 months or something like that. Right. So we can speed up one part of it, but we have to think about the funnel more broadly. We start to think about the funnel more broadly. Of course, access to incredible information on demand really does start to create speed advantage all the way through that funnel. And so when we start to think about that kind of thing and that teams start to feel the benefit of that, it's incredible how quickly you start with skepticism. Right? You lead through that moment and then you flip into, what else can you guys do for us? We're now true believers. We're on the bus. Where are we going next? Okay. So I think it's not just about that sort of initial customer listening and change process. It's about bringing your customers on the journey with you in innovation so that actually you can continue to experiment with them. Because of course, one of the biggest failures in innovation is you land innovative product and then you sit back and you go, this is good enough now.

That's right.

And of course, the problem solution fit is constantly changing. One of the things we haven't talked about is the incredible power of AI to surface data insights. If we are serving thousands, tens of thousands of responses per month from an information database, again, AI serving those on demand, the trend data we see on that, which becomes far more useful than contract metadata because it's so much more dynamic, can be queried in a much more dynamic way. You're building products with passive information that's coming to you as well as the active customer discovery side of things. Once you start to get into that level again, the problem starts to shift. You can see the problem shifting. We see that privacy and sustainability are coming in as particular issues. I'm just picking a couple of things that I know are super of interest to our customers because again, we see it in the data. Privacy starts to be something that customers are really getting excited about. And so they should be right because it's so central in building trust. Okay, fantastic. How can we start to look at our DPA and see if we can also look at the customer experience of engaging with that? How does it present? How can we shave time off that process? How can we make that process delightful for customers? How can we Shift the conversation away from the wording of the DPA towards how we're going to look after your data in a way which is like world class, best in class. And again, I think expressing yourself in those terms and also supports with that continued cultural innovation in partnership with your customers. Because again, if you're just building products in the abstract, you're not really building products that your customers want. You're building what you think that your customers want, which may or may not be correct. So bringing it back. Customers have responded to this so positively because they understand that we want them to be able to do their jobs in a way which is great. And that's a very different energy. Of course. Like, again, I reject this. Lawyers can internalize a monologue with regards to how customers talk about them and see them. I find that dialogue around those kind of things can be quite negative sometimes. You know, this is the cliches, the department of no stuff and blockers and whatever nonsense. Why would we talk about ourselves in those terms? Because part of us is sponsoring that narrative. I completely reject that. If you feel like your customers feel like that about you, you should go and talk to them and understand why. And it may be a matter of individualized outcomes. Like, I was this far away from hitting my target in a quarter and you guys dragged your feet on something and I've never forgiven you for it. Okay. Like, people are driven by incentives and sometimes those can be at a very personal level. Wouldn't you want to know? I would want to know.

Yeah.

So again, the sense that we're listening to customers as well, I think is also really important. We have some mechanisms to do that. Ultimately, it's a dialogue. Yeah.

So what makes I think you and your team such an incredible department is because I had asked you about your clients, really intending that to be like your business partners, you know, your procurement team, your product team, your CEO, which you answered. But you also brought in your end customer. Because you don't see yourselves as just a service to the internal clients. You're seeing yourself in service to the entire business objective. And that's where I think the in house teams are really changing. And as a firm, we have to pay attention. Right. So you're no longer just a responsive, get contracts done, get out of the way. You're thinking about the entire life cycle of your own travel perk customers and how the legal department is viewed from their perspective in addition to your internal clients. And then as a firm, we have to think about all those things that are changing in the way Legal teams are working in house too. It's, it's fun. Yeah.

Yeah. Well, I think it's. I don't want to get too existential. Right. But why are we all here? Don't get me wrong, we still spend because we're in a pretty active business. We're fundraising with, we're buying things like, and so on. We still spend a good chunk of money on external council. I treat that as an investment in travel, perk winning. But I find myself increasingly intolerant of, of not necessarily inefficiency, but very rigorous thinking about how things are going to be done. So that picks up a theme that we touched on earlier on. But I think this is also part of the transition. I know think about clients. One of the things I think that law firms are probably trending on the wrong side of is people take sort of, what's the word? Security, safety. They feel like it's more reliable if things are done the way that things have always been done. And I think there is a cohort of clients who actively want to be worked with on something which is different. And I would put us in that category. Right. So as much as we're working with Start, like I work as much with startups as I do with law firm partners and I see the tensions on both and I know who is closer to probably getting towards a long term solution. Right. So I know you talk with some of your other guests around, you know, legal as a service concepts and those kind of things. I think that's absolutely right. But law firms shouldn't give up on the idea that they're incapable of innovating. Just as anybody listening to this who may be struggling to make that shift from a culture of which is quite staid and secure, something which is a little bit more innovative, it's all possible. But I think part of it is acknowledging some of your own blind spots in terms of how you approach these issues and maybe just a little bit being a little bit less, a little bit less arrogant, a little bit more humble about what you do and don't know about what it is that your customers want. I do think that is a really critical part of this. And again, we can all get very genned up in terms of titles and seniority and in some instances pay packets. But ultimately if you're not delivering what your customer wants, they're going to go somewhere else. So with in house, that somewhere else could be today a Google search, tomorrow a Gemini search, and maybe they like what they get and maybe they don't come and talk to you anymore, which is worse than them talking badly about you.

Yeah, it's the classic innovators dilemma that I think it is exactly that the law firms are sitting in. We've been very successful. This model hasn't been broken yet. But the headwinds or the existential threats are coming. And do we have any of us have what it takes to pivot to the direction that our clients need us to go? And there's plenty of competitors, be IT alternative service, Big 4 technology in house teams themselves that can take a piece of that pie from the law firms. And you know, it's been happening already. So it's really up to the firms to kind of figure out how do we change such a giant, giant organization and culture and successful model and take a pretty big risk or take a big chance. And I have said this since I was at a firm 20 years ago, it's just going to take one, one of the top AMLO firms to pivot and do it successfully. And then every, I mean, we see this all the time, right? Everyone follows immediately. They copycat. We see this with, you know, associate bonuses or rates or new office in this country. Like one does it and everyone will do it. Who's gonna be the one who is willing to take that risk and is successful in it. And there's gonna be a lot of risk taking that's not successful. So not everyone will be around it. It's a fascinating time.

Yeah. I mean, if anybody's not read the book the Innovator's Dilemma by Claydon Christensen, you're gonna learn a lot about random different sizes of hard drive when you read it. Like first physical disk drive, which dates it straight out of the gate. And also a little bit about steel mills and mini mills and so on. But one of the points there is initially, essentially, innovators create a new market that eventually eats the main market. And one of the classic mistakes incumbents make is that they dismiss the technical capabilities of the innovators initially when the innovators improve. So they build this market that is not capable of being served by the incumbents at the right margin or it seem to be too small. They build that market. They do it very efficiently, and eventually as their technical capabilities improve, they eat the bigger market. One of the examples here is in steel mills, the idea that you launch this thing called a mini mill, which initially created very poor quality steel, but eventually the technology became good enough for it to create great steel and it could do it at a really good margin. And that's when big steel mills started to close. I think, you know, I'm sure anybody listening could just take the obvious lead here. I think the interesting question is at what point does it tip over from. You can actually use the brand equity that you've built as a huge law firm to sponsor your own disruption. So you move into the space where you serve a market right now which is a little bit too small. It's not your 2 billion, everyone, all the partners get a yacht kind of situation right now. But in the long run, it's worth placing that bet whilst your brand still has equity. The moment that somebody else's brand has grown and seems to be that Challenger brand, right, that's when you end up being Kodak. Now, again, one of the key observations of that book is we think about the Kodak and BlackBerry and these are the kind of companies that have very famously failed to pivot. We think about those companies as being dumb, but actually they just became too good at serving their existing customer base. Everybody told BlackBerry that they would only buy a phone that had a keyboard attached to it. And I was the BlackBerry person. I'm going to guess though, that you were a BlackBerry person. I love that thing, like the Q10, my favorite phone ever. But. And they would have said to me, I have to have the keyboard because I'm a business user. It needs to be robust and, you know, I need to be able to take it anywhere and blah, blah, blah. And then someone comes along with something completely different and better and then it kills it flatly dead. That was only just because BlackBerry listened to their customers a little bit too much and didn't think about the customers that they weren't servicing at the margin that their investors wanted. When companies fail to pivot, it's just because they do what business schools tell you to do, which is to continually refine your product to make it better and better for the customers that pay you the most money. We're focusing here on law firms, but I would also say in house, one of the, in my view, mistakes that GCS can make is to build a product that works only for the C suite so you can come in and go. And again, this is where you see dashboards coming, which I am very anti dashboards. If you have a product that includes a dashboard, I do not want your product. Please do not contact me on LinkedIn. I'm not interested in dashboards. Unfortunately. Now that will result in a lot of people saying, coming into my LinkedIn and saying, I don't have a dashboard, or I do Have a dashboard. But I'd love to show you why you're wrong about dashboards. I don't want to talk about dashboards. But what are we trying to show with that kind of thing? Sometimes it's like the illusion of control, which is something that lawyers really love to have. Like, I'm across everything. I encourage everybody to release that idea, but also so you can go in and go look at what's happening in my department, you're signaling value through pie charts and graphs and all these other kinds of things. Unless you're making decisions based on that data, it's actually informing the day to day of how you build a product. That data is useless. And so you shouldn't put any time into showing off to the people upstairs because there's five to 15 of those people. Whereas in within Travelclick, I turn to the rest of the business, where there are thousands of people. You have needs that, if we solve them at scale, drive the numbers in an exponentially different way. From an outcome standpoint, that's your opportunity. It's a messier and naughty group of people and they have different views and conflicting views and so on. And that's customer discovery at scale. It's hard, but great ideas come from. More likely to come from the coal face. And again, I'm sure yourself, when you talk to new people that join the firm who have done things in a different way and have a different perspective at any level, junior people that joined the firm and are coming into law now and are looking at what's in front of them and asking questions about why things are done the way that they're done. So much value in those perspectives. But again, if you're in an incumbent mindset, it's much harder for you. You're not an expert on this, you're not a power buyer of phones with keyboards or rolls of film or whatever it is, those perspectives are the things that drive that wonderful innovation. So again, part of running a team that is creative, that sense of humility to put yourself not at the top, I have to be the person with all the ideas, but actually to be in a position of being humbled by your own team's perspectives. I mean, the quality of when we see young lawyers coming in or even people that we've brought from outside of law in, and then we're putting through the SQE route here in England or Wales or whatever, the quality of first principles, thinking, curiosity, these things that we hire for, it's just. It's like next level stuff. And I love that perspective, it's so enlivening to me. It's so refreshing to me. And again, it's been an incredible source of gains for us as a team to be able to capture that perspective and marry those personalities with other great people in the business who think the same way.

Yeah, I have loved that. When I've built my teams in the past, I have, you know, always built my legal operations teams of people who are not lawyers or former paralegals or people who've come from anywhere near the legal space. Because I loved that, you know, perspective that they brought in. And, you know, people would always come into a new environment and say, oh, you know, this is so different. I don't want to ask dumb questions, but like, how does it work in this legal space? And I encourage the dumb questions because they are fundamentally challenging things that we take as just truth. And that's the way it is in this world. And when someone from the outside goes, that makes no sense. This is crazy. You're the only industry that does this way. You know, we have to have more people thinking like that.

Yeah, everybody listening to this recording, go ask a dumb question today. You never know what you might find out if you ask a dumb question. But it's super true. The baggage that these lawyers can have around needing to be perfect, needing to have the answers to all the questions, to be seen to be that person. Take the pressure off yourself.

Yeah.

Particularly when we're thinking about landing technology. Because I can tell you this, it's not going to be right first time.

Totally.

And if you invest yourself fully into it rather than again thinking about landing and iterating, if you spend X amount on the first version of it, but from that you're prepared to be, you go, okay, it didn't work. But now we know what we're actually going to be buying better or we're able to integrate it better, or the customer experience needs to change. So we're going out with a different set of criteria when we're shopping. Those are all wins. Go out beating yourself up, acknowledging again, my privilege and being able to do those kind of things.

No, that's exactly right. I mean, I always took it for granted because I was at Google for such a long time where everything was innovative and everything was new. And so that was just part of the lifeblood and the culture there. And there was such a focus on psychological safety. I mean, I think the Google people team are the ones who actually brought that up as, you know, what makes a great manager, right. Is that's the number one criteria. You create a sense of psychological safety. And that is so important, probably more so in the legal space than anywhere else because of what you just said.

Right.

There is this training desire need to be the person that points out the one thing that could go wrong. And that's what makes them great at their jobs. Right. And we need them to keep doing that in their day job. But in a world where we need to think about what does tomorrow look like, it is going to be about risk taking and doing a lot of things that don't work. And to your point, you know, I sort of always hesitate with the we love failure culture in Silicon Valley because that's. That is not what this is about. It is taking risks, but being very agile and knowing quickly how to pivot and how to be flexible and how to be comfortable with. Oops. Well, that's not what I said was going to be how things looked in six months, but now, six months from now is going to look different because all these new things that we learned or the world has changed.

But again, part of that is, and this is going to be an area I think the lawyers will need to upskill, but also to realize the potential of the data insights to come back to that subject. Right. Again, we know that AI has an incredible power to crunch through information, produce observations on that information, and at speed, we move away from the idea of that being just solely something that brings efficiency to discovery, for example, which is obviously a use case for AI, and start to think about that. What other data sets do we sit on that could be leveraged to the benefit of our customers? I think that's a whole different ball game then, particularly when we marry that with. And again, this is not about diminishing humanity. Okay, I know that's an obvious thing, but for me, data insights to open up conversations that are had by real people that then result in actionable outcomes. That's the pathway. And when we think about the value proposition, you must get frustrated. I get frustrated by this, where we go to an event and you hear people say law's got to change. And everyone goes, yes, law really must change. And don't we all agree that law must change? And everyone goes, yes, law must change. End of panel. Okay, so what if all deliverables are going to be created by AI, like on demand, within the near future? Let's call that a year away, probably tops. Okay, we can't make about the deliverables, we can't make about how quickly they got the phone. Can't make about how quickly we respond to an email. All these things again being proxies for value rather than value itself. Okay, this is a two sided thing. Some things will get dialed down, some things that we used to be able to charge for or used to be able to spend time on or were seen as wins for us are going to get diminished in value. But other things that were not possible before are now possible. And again for me, starting with these data insights. So why is the CFO the right hand of the CEO? Because he or she has the core numbers that tells the CEO how the business works. That is a single source of truth. Okay, what's our version of that? Can't be contract metadata. It helps us optimize our product, it allows us to create insights. And you know, if we've got a diligence round or we're raising money or whatever it might be like, it's nice to have this idea. We can dive into our content set and have ownership on it. Okay, what's the next layer of insight that we want to create? If we start to think about what can we bring to the party which is unique? I think part of that again is not just these kind of what's happening within the business right now. What conversations are being asked of models in the business right now, which is technology, which is currently available. Everyone on this podcast could go and open that up. What else can we start to bring in that allows us to bring a behavioral insight into perspective? So not just the neutral numbers, what's actually happening in the business right now? Where are the pain points in the business right now? A lot of people not knowing the business right now, that's stopping them from working. Because just like finance, legal sits across the whole of the business in the past we've needed manpower to go and locate those kind of things, like ears to the ground listening to that kind of stuff. Now we have two things. We've got data insights and we've got that so we can humanize what is fundamentally to some people might be a bit of a weird thing like this AI is telling on us the entire time. People are asking lots of questions of it positively it could know that we'd like to know more about sustainability. On the other side, it could be there's actually some whistleblowing stuff going on over here that we need to know about. But again, either way, really important still things that I need to go in there and dive in. Okay, so I know you're asking that question, but what do you really mean why is this important? That, you know, this thing, are you seeing this anywhere else? And I think that combination of our own innate nosiness as lawyers, I necessarily what it is, we tend to be quite nosy people bringing that nosiness. That's not just on a hey, legal's here and now we are. Okay, so how can I help? Or maybe you could make that contract go faster. Actually, the conversation is initiated by a data insight that we were able to generate and it becomes so much more of a fuller conversation and we can really get to the question of creating wins for the business, be that commercial wins, driving the top line or bottom line, or addressing things that could impact the top bottom line were they not addressed. You know, getting our arms around things early. I see so much possibility with regards to the management of behavioral risk in a way which is just completely next level. This is the promise of big data from the early noughties. But it's so much more actionable than that.

Yeah, exactly. Right, let's get tactical and practical and real for a second because I think so much of what you guys are doing is interesting and I've always said the real innovation is always going to come from the in house legal teams who are doing things that are forward thinking. So how are you guys practically using AI and tools differently than you were, let's say before 23?

Well, to start with the same, which is that AI driven chatbots, which is the cliche of AI deployment work, they offer solutions at scale, they do it in a way which is predictive. If you were to imagine a chart with volume of query and complexity of query as your axis. Right. The high volume, low complexity stuff, you don't go to law school to respond to that. So let an AI take care of it. Again, something which is available at a very reasonable price point. What's changed, I think is that obviously the quality of the models is changing. And one of the things that we're seeing, I think there's probably two angles to it. First of all, it's becoming way easier to build your own solutions within models. The models can actually coach you into the process. So we've done some interesting stuff with the current iteration of Claude, which again is not a particularly expensive thing. But I think it takes a certain type of culture to be able to create the space for someone to go and play with it as against the customer problem, particularly in circumstances where maybe a simpler way to solve that problem exists now. But we want to try and work it through a model to try and create a different level of understanding of it. We're also seeing just out of the box models being better and better at not just creating agreements, but also starting to mark agreements up. There are obviously again a lot of startups working in the marking documents up space. What we want to do, very much like with knowledge or contract automation, is to get completely out of the way of that process. Again, we're seeing a higher level of discretion in the markup, by which I mean early versions of this would just delete an entire clause or like completely rewrite a document paste over the top of it. That's not how lawyers actually draft. As the quality of the underlying model improves, the quality of the markup products improves, as the quality of the market products improves. It's another sort of bucket of things that essentially we can just get completely out of the way of. We automate as much of our own paper as we possibly can. So we think about 7,000 ironclad workflows last year. That's just we are not involved in any of those things. Any is not quite probably putting it a bit strongly, but a very small number and certainly I'm not involved in any of them now. We want to start addressing the same or create the same level of automation for third party paper, and particularly to create ways for our customers to be able to essentially run through a process whereby if we can't at least automate the first, like entirely automate the markup, we provide a really delightful and predictable experience on what happens after that and very gradually just continue to eat into the space of things that can be delivered conveniently immediately, visibility on demand, et cetera, like I said in that earlier keynote. So it's all about getting ourselves out of the way of the process and it's all about allowing our customers to work at the timeline which is most suitable and convenient for them. Once we build the pathway for people to use those solutions, then we can start to innovate on that. So obviously AI can work in lots of different languages, which is very helpful in the multilingual business, like Travelpook. But also can we start to build coaching tools? So we start to bring in use different agents, for example. So one of the applications that we've built since that keynote was literally a coach for people to explain how to respond to responses from customers. So when a customer comes in with this kind of query, what is it they're really asking? Because it's not always clear. If you get something, particularly from a legal or a procurement team, it might come couched in jargon, it might refer to a particular clause you might not have the document in front of you. So how can we help people get a really quick understanding of those things and then help them understand how best to respond in a way that works at TravelBook, irrespective of whether or not we're online, offline, busy helping people to work at the speed that they want to work. Which is a great, again, a great accountability driver for TravelBook because it can be very easy again for customers to say, oh, the markup's with you guys. Reducing that contract velocity is a deal killer. That's not something where we're saying, okay, we'll just do the contract faster. We're just trying to think all the time about that next level. So agentic AI, specific agents do specific jobs is also an area we're exploring. But also having in mind making these things as user friendly and immediate for customers because again, we can build very technically clever products, but if the customers don't use them and we derive no value from them, that's a waste of time. So sometimes we have to remind ourselves to again go back to that problem solution fit. Are we still solving for the right thing? That product might have worked then, but is it going to work now in this different form? And we've had to. Your point about failure, We've had a lot of fails, but it's okay because our internal clients know we will try anything to help them be better in a way that works within Travelpost findings.

Think about the risk tolerance level, right? And you're empowering your end customers or your internal clients to do some of the stuff self service or to work with the AI itself that you have trained and you're moving up the value chain. There's going to be a point where there is some risk and it could get something wrong. But you're comfortable with that, which I get because hey, you're saving tons of time and it's worth the risk reward trade off. Where do you get comfortable with it's.

Okay, so on the chatbot side, it's only referencing materials that we've taught it on. And again, there's tons of products in this space. Essentially you build a box, you put into it what you want the AI to report against and then you put the output channel so it could be slack or teams or whatever, and then you go from there. I think we have very high confidence in that. And we still, we're still clearly in those threads. We're not gatekeeping them in a sense of okay, I have to clear absolutely every one of these responses and because our customers are our partners in building these products. We're a slack business, you see in channels. Occasionally the bot may come up with something that's out of date. So let's say there's been, perhaps there's a change to our board or we move offices in one of our locations.

Yeah, perfect example.

Someone will spot it and they will tell us, we will go to the underlying materials and we will adjust it. And you can assume, I think, that essentially models won't become self regulating, but certainly if these things are going to scale to suit enterprise needs more broadly. So the UX will need to evolve. It's not just that I have to go in and address the underlying document, which is still the better way to do it, rather than trying to fine tune, for example, putting another layer of knowledge on top of the materials. You've already taught the model there will be a way to regularly and quickly change those kind of things. Again, not thinking about it as a risk, we're thinking about as an opportunity, certainly for law firms, once you start to scale out that kind of internal knowledge base. And it can be very quickly and rapidly iterated by users in a way that enhances the model for the benefit of all users. Because again, here, these are all scale tools, these are all force projection tools, these are all things where, if we think about something like the mobile phone, for example, it increases your availability, you can be contacted during more hours in the day. But ultimately there's still a cap on how long you can spend on a mobile phone in a particular day, which I'm sure you've tested the limits of on certain days. But actually here we're talking about that piece of knowledge now goes in and it's available to the whole community immediately. Yeah, right. So again, the power of that is so substantial, we continue to play with things, we continue to speak to lots of startups in this space, all of whom are delightful. Like, I love speaking with product innovators and ultimately I work for product business. And as much as I see some staidness in how we approach things, like we are an inherently conservative profession, so that's completely okay. I am so excited by the innovation in this space and I prefer to see this less as being a what are we going to lose as a profession, which is still there all changes. Loss management is so true on this particular issue and I would so much rather focus on what the future state of this could be. Because when we think about things like access to justice, when we think about even just serving the needs of a Younger consumer at scale, the possibilities are just so vast. I encourage people to position themselves more in that space. Think less about what you're turning your back on and what you're turning your front towards. You know? You know what I mean?

Yeah. On that note, kind of the most common question I get when we think about trying something new is how will we train the associates? How will, you know, all these young folks learn if they're not churning through documents, negotiating NDAs, going through the pain that many of you went through in your early days of training, what are your thoughts on that?

I don't have a great answer on that question. I think that those things will gradually become less important. Models will regress towards a mean. When I hear that question being put, I always think of a friend of mine who was very senior, a very large firm, and had kind of an epiphany Sunday morning, 3am when they were negotiating agreement with another large firm and they'd been arguing for three hours about whose entire agreement clause they should use. So zero value being delivered to the client. Because no, if I'm a client, answered situation, and I heard that you've just spent three hours argument, entire agreement clause, you better believe I'm not paying for the last three hours. But that's beside the point. It's more about what does value really look like and how do we train people to deliver value. I get that there needs to be a technical core to this, and I do also see the risk of essentially a winnowing away of not necessarily technical skills. But the more we just rely on models to do things for us, there is a sort of a timeline where you end up with, you know, the. Whatever it is, the Pixar animation with the little robots where we're all just like sitting in chairs, thank you very much. With our big thumbs and big guts. Certainly we should be thoughtful about that. And we've seen that once people move from knowledge retainers to knowledge indexers, they become very reliant on models to be accurate. So actually the pressure for models to be accurate goes up.

Right.

But again, what I would say is it doesn't stop us from leaning into actually what the problem to be solved is, which is generation of value, solving a customer problem in a way that the customer finds delightful. And I don't think necessarily that the old ways of doing things where it was completely okay to do that kind of thing. Right. Just crazy levels of negotiation on things that just require, at least on a risk perspective, very limited negotiation. That's not a behavior we want to embed. And so what does training look like? I'm not also someone who thinks that all lawyers should be able to code or something of that kind of nature. Like, I can drive a car perfectly competently. I couldn't take one apart with a spanner. But again, I think it's probably thinking a little bit more broadly about the kind of skills that would inform a value proposition in the future based on certain assumptions. So if one of the assumptions is all deliverables are available on demand, we are a massive firm. We've got a huge history of delivering this kind of stuff, which means we've got a huge amount of data, which means that we can predict what clauses have what impact in the long run. Okay, how much negotiation do we need to do? Or is it becoming more of a model versus model and thinking about how the data insights that we have can start to much better inform the value that we deliver and what skill set is required to bring that to bear. Yeah, so very pleased to be talking with a couple of educational institutions on this very subject because these are going one level before we go into the law firms. What are we teaching people at law schools? Like, certainly my own law school experience was that it was a absolute waste of time in the most possible way. Apologies to the law society for ticking all their boxes and paying money to do it, but there we go. A delightful year. Because I knew that I was going to get absolutely murdered when I went into my training contract. So very good to just be chilling, but at the same time not something that really returned value to me. So if I'm running a law school right now, what am I thinking about educating the next generation of lawyers to be ready for an environment where, okay, there's now multiple pathways. Obviously, lawyer, you test, note yourself. There's loads of ways to innovate in this space. How am I training that generation? And I think you again, use something you've searched on before. Would you have said to a son or daughter or niece or nephew or whatever, younger sibling who wants to go into law, Would you have said, yeah, definitely, go for it 10 years ago, probably not now. I would actually be more positive towards that kind of thing. But it kind of tracks all the way through. I don't. In the nicest possible way. I want lawyers to be very successful. It's like an accelerationist approach. We have to accelerate into a space where actually we law firms don't know what to do with you guys. And people take their incredible talent, dedication, hard work, like all these great things that Lawyers over indexing, client sensitivity, complete a finisher sort of mindset, like all that kind of stuff, and they redirect it to the future state of law.

Yeah.

Then that's a good outcome. I know that there may be some casualties along the way and it's very easy for me to say that as somebody who's general counsel rather than somebody who's starting out in their career, but if law firms don't have the answers to those questions, people are going to go and find a different way of doing it. And maybe that might not be too bad a thing. Again, accepting that's going to involve quite some disruption.

Yeah, I think you're exactly right. I mean, thinking back all these years, people have been saying what you've said forever, which is, I didn't get the training that I needed in law school. And so you get it at a law firm. And the in house teams aren't ready to hire young associates out of school because they're not trained. And the in house teams aren't in a good place to be, you know, conducting those in depth long mentorship and training programs. And so it falls onto the law firms. And does that make any sense? Coming from any other industry? Does that make a lot of sense? The whole ecosystem needs to change at the same time. It's a fascinating time. Andy, as we kind of wrap up, we're in January. What does 2025 look like for you?

Goodness me. Well, one of the challenges of having a very curious and innovative team who has loads of great ideas is you don't starve for ideas, you drown, as they say. Right. We have various processes to try and shake ideas out and then ensure they prioritize against customer needs. And so obviously we've done a lot of discovery as we've led into the new year. Right. So we do a lot of internal goal setting. There's a lot of planning going on towards the end of the year. But actually what I really want to do is we hired a lot of great new lawyers last year. We created opportunities as well within the business for people to take different pathways. What I really want to do is delegate hard to my team, free up my most senior people to bring their creativity to bear. Because there's also a danger here that you kind of crush a middle management layer almost into you have to manage the innovation for the people below you, so to kind of create opportunities, but you don't always get to be innovative yourself. And I think, you know a few of the characters in that layer and they have incredible capacity to surprise and delight in terms of their own creativity and innovation. And so I want to make sure that everybody is able to deploy their full creativity in response to customer needs. And so a lot of the questions that we're asking the team at the moment are along the lines of, what's stopping you? What are you perceiving is stopping you from doing something really radically different. And clearly I touched earlier on on models evolving. You'll also be familiar as kind of like a predicted bonfire of enterprise tools because of innovation of AI. I think one of the big questions for us is when do we stop tinkering around the edges with this and we actually build with AI at the center of the stack from start? That is certainly a process that we're going to be looking at this year to just sort of. To fling a little bit with it and really jump into the. To the stormy waters. And I think there's huge enthusiasm for the team for where that could take us in terms of the experience. And I think a complete revolution of the experience is obviously scary, but I love part of the culture of being building an invited theme is to say to people, if it's scary, it's good.

That's right.

If you're embracing this change of scene as a world of possibility, then it's game time. And I'm excited by that. But that's not going to happen. If anybody feels any constraints on the ideas that they're able to bring to the table their own process of innovation. Tom Rice has a great phrase, which is, this is not weighed down by the present when we think about innovation, and I think that's absolutely right. I don't want us to think about what we need to do now. I want us to think about what travel needs in 2027 and beyond. I think if we can get ourselves in that space, which I have confidence that we will, I think we'll be able to continue to do some super exciting, fun things.

Yeah. Amazing, Andy, thank you. You will. And continue to be one of my favorite gcs out there. Always fun to talk to you. Such an inspiration.

Oh, thank you, Mary. And thank you again for having me on the podcast. Lovely to see you again, and I hope we can touch base in real life very soon.

Sounds great. All right, everyone, thanks for listening.