Andrew Cooke is Chief Legal Officer at Perk, the leading AI-powered travel and spend platform. Perk's legal and privacy team has been recognised by FT Innovative Lawyers and The Lawyer for its pioneering work in the use of AI. Team members play leading roles in AI working groups and legal communities across Europe and the US. Andrew speaks regularly to diverse audiences regarding technology, status anxiety, and legal as a product.
You've worked across entertainment, esports and technology, including eight years in Abu Dhabi. How have those experiences shaped the way you think about legal leadership today?
Building effective culture—and particularly, giving teams a sense of belief, purpose and mission—is sector agnostic. Meta-research, like Gallup’s Q12, supports this view.
One of the great unlocks for any legal leader is to de-condition themselves from behaviours of private practice—where management capability is typically assessed via the single lagging metric of billing, rather than on helping individuals to fulfil their potential more broadly—and start to pay more attention to the fundamentals of engagement: from the basic needs of expectations and resource, through encouragement and recognition, to belonging and purpose, up to full realisation of potential. My time at INSEAD was a tremendous catalyst for this thinking, in that it helped me better understand how seriously many (or perhaps most) successful organisations take the practice and development of management skill.
You describe change, demystification and happiness as the DNA of your legal team at Perk. It's an unusual combination. How does that translate into the way you lead?
In building culture, I find it useful to have a simple touchstone phrase which you can (a) return to as a directional line-check and (b) through its simplicity, hammer into your team. Perk is a company where change is inevitable; so having change first speaks both to that inevitability and our role as agents of change. In our view, too many lawyers hide behind (or seek to derive power and status from) complexity—so the word demystification is a reminder that we exist to do the opposite. And then happiness speaks both to customer happiness—which is our north star—and the right of the people who work in our team to derive satisfaction from the work they do.
Your article "The Legal Ego Must Die For Law To Be Reborn" challenges some deeply held assumptions about the profession. What prompted you to write it, and what reaction has surprised you most?
I think a lot about problem <> solution fit in legal—as in, are we certain that the service we’re providing meets the customer’s needs? Do we understand what those needs are, and how they might be evolving? Through working in fast-growth tech—which is inherently risk-forward, and less inclined to accept convention in task execution—I’m forced to confront these questions more often than, say, a lawyer working in a more conservative business, or a business with long established working patterns between in-house teams and their internal clients. (I see this as a privilege rather than a problem.)
I also observe that there seems to be a rush among lawyers to adopt AI as an end. I see this as a function of status anxiety. Broadly, lawyers want to be seen to be on the right side of this technology (because there are clear personal branding benefits from this), and I sense this is influencing the procurement decisions they make, and the tone they are setting for their teams.
Any lawyer who adopts an AI tool as a status-protection measure will likely fail to grapple with the wider implications of AI for problem <> solution fit. For example: LLMs can create any legal work product, more or less immediately, but certainly very confidently, for the cost of a few tokens. Should we respond to that reality by tutting about LLMs’ current accuracy (which is the worst it will ever be)? Or by telling ourselves reassuring lies to the effect that, while simple work is easily automated, LLMs will never manage the very complex work I do?
I think not. We are not immune from disruption (and indeed there are many people who would like to see us disrupted). We are not immune from wider trends in how services are consumed—and this includes confronting the level of friction we introduce to a process. Our customers have a view: we should listen to it with curiosity and humility.
As to the reaction: the negative responses fell into two camps. The first was that I was giving a stick to critics with which to beat up legal. That was surprising, in that it’s unduly defensive; if we are not prepared to receive constructive criticism from within the profession, we’re going to struggle with the much more emotive views of unhappy customers. The second was, broadly, that friction is acceptable where it is proportionate to the downside risk. This point has greater merit, with some caveats—the main caveat being proportionality is usually judged by the party who benefits from the friction, so it should be tested with customers (more on this in my answer to question 6).
You make a fascinating distinction between lawyers being needed and being useful. Where do you think the profession is still getting that wrong?
First, lawyers love external validation from authority figures. They loved it at school and university, when they got celebrated for their academic achievements by their parents, guardians, and headteachers; they love the directories (which derive their power largely from the competitiveness of lawyers with regard to the rankings—I have never met a client who cares about directory rankings), and so on.
When do lawyers get most heavily validated? When they perform the ‘diving catch’. They find the technicality that kills the claim against the company. They work through the night to get the sales contract done before the quarter ends.
Why is this an issue? Because each diving catch creates three downstream problems: first, it trains the customer to understand the role of legal as being a clean-up crew for the externalities of poor business behaviour; second, by absorbing the consequences of those externalities, it creates moral hazard and makes it more likely that the behaviour will repeat; and finally, it creates a perverse incentive for the lawyer to create a crisis, because it is only by resolution of a crisis that the lawyer receives validation.
Furthermore, the behaviours of lawyers who come up through the traditional private practice route reflect their incentives. To give a simple example: you must insert yourself between the client and the outcome, because gating the outcome is an essential feature of a system whose purpose is to generate fees. This can lead to the (incorrect) belief that clients want lawyers involved in legal outcomes, or indeed that adding a lawyer to a process is always a net benefit.
Again, we must be careful to ensure that the stories we tell ourselves are grounded in truth. A story of neediness is “The sales teams love me because I always work through the night at quarter end”. A story of utility is “I never have to work through the night because I built a workflow that automates sales contracting”. AI will make it impossible to cling on to the neediness narrative - you are entering an arm-wrestling match with The Terminator. The only path forward is utility.
You've argued that AI is challenging lawyers' professional identity, not simply their technical expertise. What does that mean for the way we develop and lead the next generation of lawyers?
The correct answer is: no one knows for certain. Based on the experiences of our junior team, my perception is that training pathways (like the SQE and training contracts) were not keeping pace with technology and customer expectations even before AI. That gives little reason to think they will cope now.
I go back to a general theme of these responses, which is that AI is not only a technological development, but also a vector for a more significant cultural shift. The clock is ticking for any practice that considers ‘accurate deliverables’ to be its principal output. There are businesses operating now that obtain legal work product from LLMs exclusively; and it is important not to be Kodak, who believed its customers paid for crystal clear film images, and therefore blinded itself to the possibilities of digital photography on the basis that the photos would never be as clear as film (without understanding that the clarity of digital photo technology would continually improve).
To arm junior lawyers for the future is to step away from the comforting certainties of the past (and, probably, from the habit of lawyers of my generation, myself included, determining what is the ‘right’ way for a junior lawyer to train in an AI world). To that end: SQE1 is largely a test of memory and endurance. Are those the essential features of a lawyer who must compete with an LLM that has instant access to most of what has ever been published online and is available 24/7?
Similarly, some law firms are blocking trainees and junior counsel from using frontier technologies, based on procurement and confidentiality rules that may be over-zealous (such that it can be impossible to have a call with some firms that isn’t on Teams). This is leaving money on the table: wouldn’t it be far better (and drive a greater sense of mission) to harness the creative potential and AI-native status of young lawyers to explore future revenue streams and novel delivery models? The answer is not to shield juniors from the tools, but to be deliberate about which struggle we keep (see growth friction in my answer to question 6): rote recall goes; practising judgement, and being made to explain why the AI is wrong, stays.
In your follow-up article, you make the case for preserving certain kinds of friction. How do you decide what to remove and what needs to remain?
I draw a distinction between friction that exists for good reasons (and should be kept/defended, i.e. “Good Friction”) and friction which exists principally to protect lawyers (and should be ditched, i.e. "Bad Friction").
I see two types of Good Friction. The first is growth friction, which is the friction of repetition and struggle, and of demonstrating adequate depth of understanding before a step is allowed to be completed. Growth friction builds the lawyer and trust in the lawyer. Without growth friction, and with systematic offloading of origination to AI, the knowledge needed to assess and improve legal systems intelligently—to know when the AI is confidently wrong, to spot the edge case the training data never saw—quietly drains from the profession.
The second type is moral friction. This is the tension that arises when what the business is about to do conflicts with the values it has itself professed, or with the rules that bind us. It is the friction of saying the unwelcome thing to the CEO. Moral friction derives not from the lawyer’s personal ethics imposed on the client. It derives instead from (1) perceiving and vocalising the client’s own stated values and best interests; (2) ensuring those values and interests are applied with clarity and consistency; and (3) honouring professional obligations, i.e. the rules and duties that make us lawyers, rather than people who happen to know the law.
We should not hack blindly at friction, in pursuit of mere efficiency. Instead, we should discriminate between Bad Friction and Good Friction, root out Bad Friction remorselessly; and defend Good Friction from first principles, accepting that a legal profession that is fast, accessible, and morally absent does not represent progress.
You've written about moving from traditional legal work towards judgement, design and systems thinking. What have you actually changed at Perk that reflects this approach?
I was hired to build full sales contract automation for Perk. Before we delivered that automation, 100% of sales contracts were generated manually, and almost every one of them was redlined. Today, roughly 3% are touched by a human at any stage of the process to signature, and the scale of our output is roughly 40× what it was 5 years ago on a flat headcount.
This outcome is not merely a technology implementation. It results from (among many things) an intense focus on customer experience (and particularly delivering outcomes for our customers in the tools and platforms they already use, rather than forcing them into the platforms we control), a series of subtle changes to the design of underlying agreements, the gradual abandoning of negotiation playbooks (in favour of simpler, customer-friendly positions that require no adaptation), and the application of insights from metadata to understand the issues to which customers pay attention. Crucially, it also reflects the situation on the ground: we have one lawyer for every 120 sales colleagues. Getting lawyers out of the way was less a philosophy than a recognition of structural arithmetic.
There's a strong human dimension to your writing about change. Why do you think organisations so often overlook the emotional side of transformation?
Thank you for recognising that. Going back to our DNA: I really love lawyers; I want lawyers to be happy and successful; and I want lawyers to seize the opportunity to redefine their value, rather than have redefinition imposed upon us.
Organisations overlook the emotional side of change for two reasons: because people believe that rational explanations for change are more compelling than they are; and because dealing with people is much harder than reading a nice reassuring Gantt chart.
For lawyers specifically, our identity can be a source of status, and who we are and what we do can feel inseparable. When we feel that our status is threatened, that triggers a defensive response, which usually arrives dressed as rational argument: that AI can't do it all, that AI output is categorically inferior, that client demands for speed reflect a misunderstanding of risk, that the old ways encode wisdom the new ways can't replicate. Some of these points have merit; the problem is that they are deployed as armour. It is simply inadequate to meet this emotional response with a rational response. And yet rationality is how legal leaders are making the argument for AI: it’s all metrics, efficiency and cost-reduction. The stated case is rational; the underlying motive, as I said earlier, is often status.
Your emphasis on happiness is particularly interesting for our network. What have you learned about creating an environment where people can perform at a high level and enjoy their work?
Because we spend a lot of time with other lawyers, we fall foul of a kind of sampling bias: we forget that characteristics typical in lawyers (like eloquence, intelligence, problem-solving capability and readiness to work exceptionally hard) are uncommon, particularly as a combination. If everyone around us has these skills, we assume they are ordinary. This leads us to surrender our own agency, and misunderstand our potential for impact.
If people with these uncommon qualities cannot make the most of them, and have fun doing it, then what is the point in clearing all the hurdles required to become a lawyer? I reject legal martyrdom and narcissism in all their forms. I refuse to answer “how are you?” with “busy”. I reject the pressure to be right—I prefer to focus on my blind spots, and what I don’t know. I embrace high-trust environments. I embrace giving people space and sponsorship to find what is meaningful to them about what we do.
You've spoken powerfully about rejecting legal martyrdom and creating environments where people can enjoy their work. What does that look like in your own life? Is there a particular habit or practice that helps you maintain perspective?
I doubt that anyone in my family would say that I have truly achieved balance, or the “egoless life”. I am as prone to limiting beliefs as anyone; but I have developed the habit to recognise and confront those beliefs, rather than just accept them as my incontrovertible narrative. Being “busy” is a great place to start with this habit. I guarantee that no one at work is impressed with how busy you claim to be. (They might be, for example, worried, concerned as to your ability to work efficiently, concerned that you can’t manage your time, or just amused). Busyness is not value. So it is excellent practice to consider to whom you are really speaking when you always answer the question “how are you doing” with “busy”, and why that’s always your answer.
Looking ahead, what gives you most optimism about the future of the legal profession, and what still concerns you?
What we do and how we do it is completely up for grabs. The choice is ours: we can see that as frightening (because we must now leave behind the comfortable certainty of the supply-led legal profession, where we defined how legal was consumed, and we could largely ignore the perspective of the customer), or we can see it as exciting (because our ability to create real value and utility has gone through the roof).
Since “change” is the first word in our team DNA, I’m choosing to see it as exciting; and it’s part of our mission to make a positive case for others to join us in that view. What concerns me is that we lose the Good Friction along with the Bad: the knowledge that lets us spot when the AI is confidently wrong, and the moral friction that makes a fast profession worth trusting. And I worry that status anxiety keeps too many of us arguing about accuracy while our customers move on.