Dan Hunter and Henry Goodwin look at how law firms’ talent mix must evolve as AI goes native. Do we need fighter pilots or drone operators?
AI solutions are becoming embedded in the day-to-day ways of working of law firms around the world.
Even at this early stage of AI integration, long-held assumptions around the time and people required to deliver legal outputs, as well as the pricing of those outputs, are being challenged.
Many firms have realized that they have to train their staff to use AI-based technological tools to deliver legal services efficiently and competitively and are rushing to develop various types of training solutions for their people. These approaches have ranged from introductory online programmes and bootcamps to deep dives such as masterclasses for innovation teams or hackathons to train practice groups, even as they build out new AI driven solutions. Several of the teams that have partnered with Kings College, where one of the authors has been involved in such activities, have demonstrated a high degree of curiosity and legal tech talent, adopting novel approaches to rethinking conventional legal services delivery methods. Other legal genAI training providers like Factor have built programmes aimed at corporate legal departments and have reported similar successes. There are also online providers like Hotshot Legal who offer introductory level training for firms and departments.
However, this is a fairly basic level of engagement with the AI revolution, and few firms have grappled with the reality of AI-native legal services delivery. A new breed of law firm will emerge, constructed from the ground up around AI systems, to distinguish it from traditional law firms that have AI systems bolted onto their existing technology infrastructure and delivery of legal services. These AI-native firms will look quite different from current technology-assisted firms, relying on fleets of AI agents to deliver a range of legal services in a largely autonomous manner.
This change – one that we can see already happening in other professional services arenas like programming, marketing, and HR – will disrupt the established verities of legal practice. In this essay we focus on the talent stack of the law firm in an era of agentic AI. There are of course many other difficult questions firms will have to confront, such as how to develop the necessary technology stack to deliver agentic legal services and what types of business models will work in this new world. While we’d love to talk about them, here our focus is on the evolution of talent for an AI native future.
Understanding the AI-native talent stack
The Thomson Reuters Institute’s 2025 Generative AI in Professional Services Report shows that lawyers and other professionals are increasingly positive about AI, with a preponderance (55%) of excitement and hope, as opposed to hesitation and concern. Almost two-thirds of respondents (62%) said they believe AI should be used for work, with most (89%) already thinking of specific use cases. Respondents also believe that AI will reduce costs and free up professionals for higher-value activities. Indeed, 95% of respondents said they believe that AI “…will be central to their organization’s workflow within the next five years.”
A recent report from LexisNexis found that sixty-one percent of lawyers say they are now using generative AI in their work, with a further 32% planning to do so, but with two-thirds of respondents also stating that progress is at risk of being undermined by slow organisational cultures.
That said, the transition to AI-native law is going to be uncomfortable for many firms and it’s not clear that law firms are adequately positioned to respond to the challenge. The latest AI solutions are replacing not only the software that preceded them, but also the human labour that has, until now, been required to complete the relevant legal task. This forces legal leaders to rethink the role of lawyers in an AI-enabled law firm across the entire career life cycle.
What does this rethink look like?
The next generation of law firms will probably be resourced by lawyers and technologists operating under wholly different organisational structures and business models from today’s incumbents. As AI moves from pilots to production in legal work, its impact will be uneven across seniority bands.
In the short term, senior lawyers may not be replaced, but the bar for delivering AI-literate client strategy and risk governance among the most experienced cohort of lawyers is rising rapidly. Meanwhile, mid-level associates are becoming the crucial integrators who will ensure that valuable AI-enabled outcomes become the reality. Finally, the traditional training ground for juniors is narrowing as routine work automates and this will force firms to redesign the early-career experience to optimise both internal career development and external client outcomes.
In the medium term the landscape will look very different. The last twenty years has seen a tech-driven transition from the traditional pyramid-shape[1], with a small number of partners at the top, many associates below, to a pillar-shaped model of firm recruitment that deploys technology and associates to generate leverage. This is already being evidenced in the increasing numbers of legal technologists being recruited into law firm teams, and new career paths opening up for lawyers, for example in the form of legal engineering roles being made available to trainees on qualification.
The next twenty years will see an AI-native shift from a pillar shape to a barbell shape (see Diagram 2). In this type of firm, a practice group will have senior lawyers at the top of the barbell, who find and manage client relationships. At the bottom of the barbell there will be legal engineers, who tend to the care and feeding of the AI-native systems. And in between will sit the AI models and agents, doing most of the work that until this point we have assumed that human lawyers have to do.
This change is going to have a huge effect on the way that law is practised, and the most obvious change will be to the talent stack.
Saying goodbye to Maverick
Currently, law schools and law firms teach young lawyers that their job is akin to being fighter pilots, piloting the legal process through the clouds at 35,000 feet, bringing to mind Tom Cruise as Maverick in Top Gun. But the AI-native legal future looks a little different. Rather than fighter pilots shooting through the skies, lawyers in an AI-native firm will resemble drone operators sitting in front of screens in a warehouse in Nevada making long-range decisions in conjunction with the machine. They will oversee large fleets of agents which will undertake many of the discrete parts of legal processes, and the lawyer’s job will be to ensure overall accuracy. Just as these days software developers don’t do much direct coding – AI platforms like Cursor and Lovable do the heavy lifting for them – so too will lawyers come to lean on machines which do much of the work.
And just as you hire and train drone operators differently from how you hire and train fighter pilots, we are going to see a different set of skills in the lawyers who work within the AI-native firm. This will play out in two main arenas.
Firstly, in recruitment, we will need different frameworks to choose the right people with the right dispositions and skills to reflect the way that AI-native practice groups will operate. Lawyers will still need to be capable lawyers, but they will also need to be able to work as overseers of AI agents, prompting and controlling them effectively and responsibly, just as a traditional “fighter pilot” lawyer ensures that their own work output is accurate and fit for purpose.
Secondly, we will soon face major upheaval in the law firm learning and development function. As we noted above, a number of firms have grappled with the question of how we train lawyers to become better at using generative AI. But a larger problem is looming: all firms will have to recognize that lawyers’ use of AI changes the way that they learn and, consequently, how we train the next generation of senior lawyers.
We will no longer be able to rely on the apprenticeship model where senior lawyers train junior lawyers, and junior lawyers learn through osmosis and doing. There is now significant evidence[2] that people don’t really learn from using LLMs to help them perform their work. There seems to be limited transfer learning when people use LLMs in answering questions, and, even worse, people seem to think that they are learning even when they aren’t.
This means we will have to craft specific feedback loops[3] to ensure that fee-earners can progress in the firm and take on the role of partners and senior advisors in the fullness of time. It used to be that the learning feedback loops in law firms were provided by senior associates and partners, mentoring their junior colleagues in the art of drafting and thinking like a lawyer. Since junior lawyers won’t have that same feedback loop, a new type of feedback loop, almost certainly created and run by LLMs, will form the core IP of the AI-native firm, and their development poses another challenge that legal practitioners will face in the coming years.
The AI-native future of legal practice
AI-native legal practice is going to transform the way that legal services are delivered, and create a very different paradigm for legal talent. Each of the humans in the AI loop will need a new set of skills and dispositions compared to what has previously been required in traditional legal practice. And the profession will need to approach recruitment and training in a different way to meet the AI-native environment that we will all soon live in.
Some law schools – ahem, like King’s – together with many law firms are grappling with these issues. We work on AI literacy in a range of ways, from basic classes that can be taken by all students, through to generative AI electives in the LLB and specialised LLM classes on legal tech and AI. But this only scratches the surface and it is clear that we’re all going to have to keep iterating and developing new solutions to these difficult new problems.
[1] https://journals.sagepub.com/doi/10.1177/001979399504800405
[2] https://arxiv.org/abs/2409.09047;https://documents.worldbank.org/en/publication/documents-reports/documentdetail/099548105192529324; https://www.pnas.org/doi/10.1073/pnas.2422633122;
https://arxiv.org/abs/2506.08872; Cf: https://arxiv.org/abs/2407.09975
Henry Goodwin is a TMT lawyer turned VC investor with global early stage firm Leo Capital, where he recently launched a B2B seed fund focused on the Nordic & UK markets. He has a particular interest in LegalTech, RegTech and the intersection of law, technology and business.
Dan Hunter is the Executive Dean of the Dickson Poon School of Law at King’s College London. He is an international expert in AI & law, and is the author of books on gamification, intellectual property law, and intelligent legal systems. His current research is focused on the use of innovation and technology within law, including the use of generative AI in law and the future of legal practice.
This article is also available in the special AI issue of Computers & Law, which is available to download here.
