GenAI tipping point
Last week, ChatGPT celebrated its third anniversary of its public launch on 30 November 2022. In just three years, GenAI has hit the mainstream and 2025 was the tipping point in legal, with agentic AI unicorns Harvey and Legora driving the market, raising significant investment along the way (both announced $150m funding rounds in the last quarter of 2025). Harvey reached unicorn status in 2024, two years after it was founded, and in 2025 Legora became Y Combinator’s fastest ever unicorn, achieving a $1.8bn valuation just 13 months after joining the Y Combinator programme. Clio, which achieved unicorn status in 2021, pivoted to AI first with the acquisition of legal intelligence platform vLex, achieving lawtech’s first billion-dollar deal. On the day the deal closed, Clio announced a $500m Series G funding round and a $350m debt facility, valuing the company at $5bn. Yet Harvey took less than three years to reach an $8bn valuation.
Legal tech funding reached $3.2bn in 2025, and in November, US venture capital firm The LegalTech Fund (partner Gordon Crenshaw spoke at Legal Geek in October) closed its second fund at $110m with lead investors including large law firms and established legal tech vendors.
GenAI natives
A trend that emerged in the second half of 2025 is the native AI law firm, which is defined as a firm that is built on AI, rather than dependent on AI products. Examples are Garfield Law, the first AI law firm approved by the Solicitors Regulation Authority, which automates debt recovery processes, and in the US, New York in-house legal and compliance company Norm Ai raised an additional $50m investment from Blackstone to launch its own AI native law firm, Norm Law, creating AI agents to provide bespoke legal and compliance services for Blackstone and other financial services clients. Both combine qualified lawyers with proprietary AI technology.
But transformation is not just about emerging models. AI is changing the structure of professional services. “Will the pyramid model of consulting firms survive the transition to AI?” asked an article in Tuesday’s Financial Times as McKinsey gave AI automation as the reason for freezing starting salaries for the third consecutive year. “AI disruption is more real in professional services and technology than in the rest of the economy,” commented Naaman Mian of Management Consulted.
Law firms are shifting to an AI-first back office, where Copilot is used for everything from financial and risk management to marketing and business development. As AI automates more legal tasks and processes, lawyers need AI fluency from the get-go, as firms have started introducing AI for recruitment and training. Mishcon de Reya is piloting Bright Networks’ Bright Apply screening system for graduate jobs, replacing long application forms with an initial untimed interview conducted by an AI chatbot, and Kennedys partnered with Spellbook to create AI simulations of entry-level tasks to train junior lawyers.
Firms are citing AI as a reason to cut business support teams. Irwin Mitchell abolished the litigation assistant role and Clifford Chance cut 10% of its business support team. But there are new support roles too. Linklaters launched a team of 20 specially trained AI lawyers who will work closely with fee-earners helping them develop prompts and workflows and supporting the delivery of innovative client solutions. The law firm workforce composition is shifting as business support roles and functions are replaced by AI software and expertise.
The interoperability challenge
A key factor going forward and an upcoming challenge for AI native firms and firms switching to AI first is interoperability. As Ryan McDonough, head of legal tech engineering at KPMG Law wrote last week, in order to achieve their potential, they will need to become ‘API native’ too, in order to interact with the wider legal ecosystem. Firms and legal departments are no longer looking for a stack of ‘best of breed’ technology, and most lawyers –with the exception of serious geeks – aren’t into vibe coding. Rather they favour intuitive GenAI platforms that save time, and allow them to retain control over legal processes and decisions, particularly following the Mazur High Court judgment which clarified that only authorised practitioners can conduct litigation and may have implications for other reserved activities. This goes some way to explaining why agentic platforms Harvey and Legora have taken off so fast, and AI powered interoperability is a key element of Clio’s AI first strategy.
Self-driving law?
As AI automates straightforward legal tasks and processes, freeing up lawyers for reserved activities and high-value work, it is worth considering that this is likely to reduce the number of lawyers, bringing the risk that autonomous legal services will (self) drive the sector away from its core purpose. This point was highlighted at a UnitedLex event at the Amsterdam University of Applied Sciences, where Professor Ivar Timmer showed us Viridea Productions’ AI-generated video, Barney – The End of Lawyers, Douwe Groenevelt’s dystopian extrapolation of legal’s rapid adoption of GenAI.
Professor Timmer questioned the wisdom of allowing automated decisions to replace legal argument, quoting French essayist Joseph Joubert “It is better to debate a question without settling it, than to settle a question without debating it.” This was particularly apposite given the British government’s plans to scrap most jury trials, removing an important human connection from legal process
Human connections
E-commerce leader Eva Pascoe who co-founded Cyberia, London’s first internet café where people could meet and connect online, wrote on LinkedIn, “in e-commerce, it is easy to forget a simple truth: most of us only have jobs because somewhere, someone loves someone else enough to give them a gift. Strip away the algorithms, the logistics, the dashboards, and what remains is an ancient human impulse: I was just thinking of you… As we see each other less in hybrid work so the act of gifting has become our connective tissue.” She was referring to physical gifts, but her post, and her career, highlight the importance of human connection to business and professional success. Data may be the moat, especially in relation to AI or any other technology becoming ubiquitous, but we still need human connection.
Finally, GenAI’s latest foray into uncanny valley is damaging some human connections. Last month Wired carried an article about emotional attachments to chatbots and AI companions breaking up human relationships. While ChatGPT is programmed not to tell people to break up with their partner, Wired referred to a “new legal frontier” in family law, which has been corroborated by Divorce-Online which is seeing an increase in divorce applications due to one partner becoming fixated on an AI chatbot. So just in time for January, the ‘divorce month’, AI for unmet emotional needs is creating more work for lawyers!
Legal Geek is hosting four conferences next year, learn more on our events page.
Written by Joanna Goodman, tech journalist
Photo credit (Joanna): Sam Mardon
