The Legal Ego Must Die For Law To Be Reborn

# The Legal Ego Must Die For Law To Be Reborn > On status, empathy in transitional change, optimising for utility, and the immense possibilities in the future of law [Read on Substack](https://lawwhatsnext.substack.com/p/the-legal-ego-must-die-for-law-to) · 2026-04-23 · Law What's Next --- John Philip Kemble as Coriolanus, detail of a 1798 painting by Thomas Lawrence (Public Domain/ Wikipedia) The following article was published by our friend Andrew Cooke (Chief Legal Officer at Perk) on 3 April 2026, and with his permission, we are sharing it here. Andy’s piece does something we think is genuinely important: it treats the legal ego not just as a punchline but as a real obstacle to change, and in doing so gets closer to something the legal tech conversation usually skips past - that the friction in law's encounter with AI is as much about identity as it is about capability. We should also mention that Andy is speaking at Crafty Counsel’s Frontier Legal in New York on 30 April 2026, where he will be exploring strategic risk in an unstable world. Worth catching if you’re across the pond. Over to Andy. I. A couple of weeks ago I attended an event organised by Harvey. To be absolutely transparent, I have no view on Harvey. My main takeaway from the event wasn’t even about Harvey (*). It was captured in this post. Credit where it’s due: the Harvey marketing team is elite, and appears to have an brilliant understanding of the following: Legal ego, and the obsession some lawyers have about their status The sense (consciously or subconsciously) that AI threatens that status Acting on that sense, the urgent imperative to make a ‘signature move’ that positions AI as a personal, status-enhancing tailwind That building Harvey’s reputation as the enterprise legaltech tool of choice for law firms and large companies is an essential sales feature, in that it provides downside protection for status - if Harvey fails/ is perceived to fail (which, to be clear, I am not suggesting it will) the impact of that failure will be dissipated across an elite class My overwhelming sense from conversations I had with Harvey users at the event was one of relief. We have made the big call, and we are in the lifeboat. We are safe from the rough waters that will consume those who make AI a headwind. This feeling prompts a challenge: is status obsession preventing lawyers confronting the reality of diminished product<>market fit between legal and its end user? And what does it mean to defer that confrontation? So, in discussion with Anna Guo (**), I resolved to explore the legal ego, and to take an empathetic view of the process of transitional change in law. II. This is a long one so, to summarise: Machines can do legal work. How much of it they can do is a matter for debate. But my direct experience is that, at the very least, low risk/ high volume tasks are easily and effectively automated with AI. Because of (1), AI challenges the parts of professional identity that were built around scarcity rather than service. In more practical terms: until very recently, clients had two choices: pay a lawyer, or wing it with Google. LLMs have stepped into that knowledge gap - not eliminating the need for a skilled professional, but fundamentally repositioning when, and whether, one is required. Because of (2), AI’s most profound disruption of law is not cognitive - it is social. The informational advantage that structured the lawyer-client relationship, and on which the entire status architecture of law was built, is threatened. I think this is a good thing, but clearly views differ. AI’s disruption of law does not ‘prove’ that lawyers are unnecessary. (I do not think it is an exaggeration to say that legal judgment will be vital to what comes next for society, particularly given AI’s potential to disintermediate the relationship between political power and the people affected by that power (***).) But: a significant proportion of what lawyers do is not lawyering. It is the performance of lawyering - and the market, many of whose participants have always suspected (and even resented) this, feel that they can now verify it. Dropping legal performance - and optimising for judgment, design, governance, communication, and system-level risk management - leads not to a diminished version of a legal career, but rather a more engaging and durable one. III. What is the legal ego? It is not confidence, craft, or judgment. It is the status-protective apparatus built around those things: the trusted advisor cliché, the mystique, the gatekeeping, the leverage of title, the performance of indispensability, the confusion of being a bottleneck with being valuable, leading with the expectation that your expertise is sacred and cannot be challenged. It is the part of the profession that is more interested in being needed than in being useful. Herewith, and heretofore, some examples, by way of example and not limitation: The bottleneck. “Nothing leaves without me.” “Only I can interpret this.” “The team must come through legal.” (In Shakespeare: “Let it be virtuous to be obstinate.” (****)) In an AI world, that instinct becomes pathological. The job is no longer to preserve dependence. It is to design systems that reduce dependence without reducing quality. The artisan-author. Some lawyers treat repeatable work as a stage for personal virtuosity and resist machine-generated output on principle, preferring to originate rather than assure. But if a workflow can be templatised, automated, or AI-assisted with equal or better outcomes, insisting on bespoke manual production is not craftsmanship - it is vanity. The value of a high-functioning modern lawyer moves from origination to assurance. All lawyers are comfortable being authors; fewer are comfortable being designers and editors of systems. The Chicken Little. Traditional legal culture often celebrates the person who can find the most risk. But businesses do not pay legal to merely identify risk; they pay legal to calibrate it and enable action. AI can spot issues. Human lawyers increasingly earn their keep by prioritising, contextualising, and deciding. The perfectionist. Perfectionism in law is often moralised, but some of it is fear in respectable dress: fear of being wrong, criticised, or replaced. In many legal contexts, the right standard is not perfection but appropriate reliability at appropriate speed. AI forces that trade-off into the open. The hoarder. In older legal models, the person who knew the precedent, the clause, the playbook, or the regulator had leverage. In the next model, leverage comes from building accessible knowledge systems so others can operate well without asking for permission. The smartest person in the room. This identity has always been brittle. In an AI environment, it becomes dangerous. The better identity is: the person who helps the room reach a sound decision faster, with fewer blind spots and clearer ownership. (Insert mandatory Ray Dalio “radical open mindedness“ reference.) The lawyers who will matter most are the ones who can collaborate across product, ops, finance, and engineering - and build legal into the system rather than above it. Each of these legal egos is a form of resistance: resistance to a loss of identity that the person cannot quite articulate, and the organisation has not bothered to name, let alone replace. Which means the useful (and more empathetic) question is not what lawyers are resisting, but why the resistance persists when the rational case for change is clear. For that, it helps to borrow a framework built on simple insight: what looks like resistance is usually something else entirely. IV. There are as many models for analysing resistance to transformational change (and to bring empathy towards those participating in that change) as there are consultants willing to sell them to you. One I find useful is Switch, by Chip and Dan Heath. (For me, Switch is up there with Good Strategy/ Bad Strategy and The Innovator’s Dilemma in the category of ‘most useful business books’; and Switch directly influenced how we addressed our end users in building the MVP of legal as a product (*****) at Perk.) The Heaths use a three-part metaphor. The Rider is the rational mind - analytical, planful, capable of seeing the long-term case for change. The Elephant is the emotional mind - the thing that actually determines behaviour, running on feeling, identity, and immediate experience rather than logic. The Path is the environment in which both operate. When the case for change is ambiguous, or requires weighing a complex set of trade-offs, the Rider overthinks, second-guesses, and stalls - not from obstruction, but from the exhaustion of someone given a direction but no destination. What looks like inaction is usually an unmotivated Elephant. The Elephant doesn’t move for rational reasons. It moves when it feels something - when change connects to identity, to emotion, to something it cares about. Presenting a logical case for change to an unconvinced Elephant is roughly as effective as presenting a logical case to an actual elephant. The animal will go where it wants to go. And what looks like a people problem is usually a Path problem. Put good people on a badly designed path and they will make bad decisions. Change the path - the incentives, the default behaviours, the friction in the system - and people who looked resistant suddenly aren’t. The prescription follows: direct the Rider by giving it a specific, concrete destination rather than an abstract aspiration; motivate the Elephant by connecting change to identity and feeling, not just logic; and shape the Path by redesigning the environment so that the right behaviour becomes the easy behaviour. V. As I noted in my summary: the Rider in most lawyers already acknowledges the direction of travel. They know the knowledge gap is closing. They know client tolerance for friction is falling. The change conversation in law is directed almost entirely at the Rider. It is a rational conversation; a conversation about productivity and efficiency. It is statistics, data, adoption curves. The Elephant - a proxy for the emotional mind, and the legal ego - is being neglected, and is experiencing narcissistic injury: when identity is tightly coupled to a source of status, and that source is threatened, the response is defence. The rationalisations plastering LinkedIn - 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 - are psychologically familiar. They are what people say when the thing being threatened is not their competence but their self-concept. And the Path - a proxy for the environment around the legal ego - is working against change, not for it. To my eyes, law firms and legal departments continue to be structured - in their incentives, promotion criteria, billing models, and performance reviews - to reward ego-led behaviours (such as those described above). You cannot motivate an Elephant toward new behaviours while the Path keeps rewarding the old ones. It is not hypocrisy. It is just how systems work. The anxiety visible right now - the brittleness, the credentialing, the enthusiasm for AI governance frameworks that happen to require lawyer supervision of all AI output - is the anxiety of people whose self-concept is structurally dependent on conditions that are structurally changing. In a market that is repricing legal services in real time, this is not a character observation. It is a P&L observation. VI. The important work in law today is dealing with the Elephant and redesigning the Path: changing what gets rewarded, what gets promoted, what gets celebrated, so that demystification and accessibility confer status rather than threaten it. We have to say it with our whole chest: what the legal ego protects is not the substance of legal value. It is the scaffolding around that value - the rituals, the signals of membership, the carefully maintained distance from the client - which has been confused, for a long time, with the thing itself. What must remain is real, and it is worth naming clearly. Judgment: the capacity to navigate genuine ambiguity, weigh competing interests, and make a clear recommendation that a real person can act on. Particularly, in scenarios characterised by low context, high consequence, and the need for accountability, AI cannot replace human judgment end-to-end. Trust: earned not through inaccessibility, but through being right more often than not, and honest when you are not. Communication: the skill to convey complexity clearly, move fast, and reduce friction without sacrificing the standard. These things are not threatened by AI. They are, if anything, elevated by it. A lawyer who has made this shift looks different in practice: they spend less time on origination and more on assurance; less time defending process and more on accelerating decisions; less time being the bottleneck and more time designing systems that don’t need one. They are not less lawyerly for it. They are more useful - and the profession must optimise for utility, not status. (For what it’s worth: it seems from this recent article by Harvey’s President that he agrees with me.) VII. I promise I did not write this article with Easter in mind. But, in a way that is seasonally appropriate, I’d like any reader to consider this article an invitation to embrace rebirth. As I have said many times, our DNA in Perk legal is change, demystification and happiness - meaning happiness for our customers and the exceptionally talented lawyers in our team. I feel that DNA obligates me to speak candidly about change; but also to encourage lawyers to embrace agency, and walk a different path towards a full realisation of their potential. Blending judgment, design, governance, communication, and system-level risk management leads not to a diminished version of a legal career, but rather a more engaging and durable one. In this regard, my recent participation in LegalQuants’ hackathon - where diverse legal professionals from across the world built and shipped a range of exceptional products in ten days, using nothing more than an idea, over-the-counter LLMs and $200 of Replit (******) credit - gives me new levels of confidence regarding the dawning legal enlightenment, and the immense potential of our collective repositioning. That starts with being radically honest about legal ego; and focusing our attention away from the Rider, and towards the Elephant and the Path. (*) I also want to shout out the croissant I had at the Harvey event, for two reasons: it was insanely delicious and, since I appeared to be the only person at the event eating carbs, a lot of people seriously missed out. Big up carbs. (**) Thanks to Anna, Tom Rice and Benjamin White for their input on this article. I’m very grateful to know people who think about these issues and are prepared to offer robust and constructive input, notwithstanding how busy they are. (***) Noting Good Strategy/ Bad Strategy’s advice (“A hallmark of true expertise and insight is making a complex subject understandable. A hallmark of mediocrity and bad strategy is unnecessary complexity—a flurry of fluff masking an absence of substance“), I’m going to unpack these many $10 words. Situations like this - where an AI falsely identifies a person as a shoplifter, effectively blocking that person from their local stores - seriously concern me, given the strong likelihood that governments will deploy AI to handle their interactions with the governed (on efficiency grounds - no doubt using the arguments they have rehearsed during their efforts to remove jury trials). (****) Was there a v1 of this that was structured around Coriolanus, the noble but deeply flawed hero whose extreme pride, hubris and inability to compromise inevitably lead to his banishment, betrayal, and death? Yes. Did it become clear that the metaphor was inadequate? Yes. Was I deeply reluctant to let it go? Yes. Is the irony of that failure in the context of an article about ego lost on me? No. (*****) I cover legal as a product specifically from about the 14:15 mark in this keynote. (******) Replit plus Claude is one of the most insane tooling combos I have ever used. Thanks for reading Law://WhatsNext! Subscribe for free to receive new posts and support our work.